The inter- and intra-individual variance in descending pain modulation evoked by different conditioning stimuli in healthy men

2.1 Subjects

Twelve healthy men (mean ± SEM age: 25.6 ± 1.5 years, age range: 19–38 years) participated in the current study. Besides, one subject withdrew from the experiment because he could not tolerate the cold pressor pain (CPP) (the data from this person were excluded). None of the subjects had current or previous injuries or psychiatric conditions, chronic pain or major medical conditions that could interfere with the normal somatosensory function. Informed consent was obtained from each subject. The study followed the Helsinki Declaration and was approved by the local ethics committee (VN20090047).

2.2 Experimental protocol

The subjects were sitting in a reclined position in a bed. During assessment the leg was positioned below the computer-controlled algometer probe and stabilized in a vacuum-packed kapok-filled pillow. CPP (2–4 °C) to the right hand, ischemic muscle pain (pressure of 36 kPa) to the right upper arm or tonic mechanical pressure to the craniofacial region were applied as CS after 5 min break following TS recording (baseline). Each CS was applied for 10 min, and the three different CS were applied in randomized order on different days separated by at least 2 days. The subjects were asked to rate the pain intensity of the CS continuously on a 0–10 cm electronic visual analogue scale (VAS). In order to test the intra-individual variance, the experiments were repeated on two different days (session I and II) separated by at least 2 days. As TS, pressure pain thresholds (PPT) and pressure pain tolerance (PPTol) thresholds were measured at the right masseter muscles (MAR) and left flexor carpi radialis muscle (forearm) using a hand-held pressure algometer (Somedic, Sweden). PPT and PPTol recorded from the right and left tibialis anterior (TA) were assessed by a computer-controlled pressure algometer (Aalborg University, Denmark) [24]. The mechanical pressures with pain intensities of 1.4 and 1.6 × PPT applied to the right and left TA were rated on an electronic VAS (0–10 cm). All parameters were recorded before the application of the CS, during the application of the CS and 10 min after the termination of the CS. PPTs by the hand-held algometer were determined in triplicate, and PPTs by the computer-controlled algometer were determined in duplicate. PPTol was recorded only once each time to avoid excessive stimulation and sensitization.

2.3 Conditioning stimulus (CS)

2.3.3 Mechanical pressure on head

Craniofacial conditioning pain was provoked with a specially designed compressive device (head band) (Fig. 1) [26]. Briefly, the model is based on a mechanical compressive device which is positioned over the vertex and can be fastened over the skull by four probes (10 mm radius) each separated by 90^° mounted on two clamps. A strain gauge force transducer is attached on the four probes, and pressure can be adjusted over time using the VAS feedback from the subject to maintain the pain intensity at a given level (target level, VAS5).

Fig. 1

The compressive device of experimental craniofacial pain. The device was set on the vertex. It was height-adjustable by a downwardly directed screw. Compression of the craniofacial region was achieved by tightening four, horizontally opposed clamp screws with a force transducer.

2.4 Pain ratings

The subjects rated the pain intensity of the three CS on a 0–10 electronic VAS (0 = no pain, 10 = worst pain imaginable), and the ratings were sampled and stored on a computer every 5 s. The pain ratings were recorded from the start of the CS until they returned to zero. The VAS peak pain values were used for further analysis.

2.5 Test stimulus (TS)

2.6 Statistical analysis

The VAS peak pain values were analyzed by two-way ANOVA to test the effect of CS (cold, tourniquet, head band) and the effect of session (session I and II) as the repeated factor. Absolute PPT and PPTol values at baseline were analyzed by three-way ANOVA to test the effects of CS (cold, tourniquet, head band), assessment site (masseter, forearm, right TA, left TA) and session (session I and II) as repeated factors. Absolute VAS1.4 and VAS1.6 ratings at baseline were analyzed by three-way ANOVA to test the effects of CS (cold, tourniquet, head band), assessment site (right TA, left TA) and session (session I and II) as repeated factors. Then the PPT, PPTol, VAS1.4 and VAS1.6 ratings were normalized to the baseline values. The relative changes in PPT and PPTol were analyzed by three-way ANOVA to test the effects of CS (cold, tourniquet, head band), assessment site (masseter, forearm, TA) and time (baseline, during the CS and 10 min after the CS) as repeated factors. The relative changes in VAS1.4 and VAS1.6 ratings were analyzed by two-way ANOVA to test the effects of CS (cold, tourniquet, head band) and time (baseline, during the CS and 10 min after the CS) as repeated factors. The ANOVAs were followed by the post hoc Student Newman-Keuls (SNK) tests to compensate for multiple comparisons.

The subjects who did not have any inhibitory CPM (i.e., relative decreases in PPT and PPTol values and relative increases in VAS1.4 and VAS1.6 = increases in sensitivity to test stimulus) during the CS were defined as “non-responders”. Only data from CPM-responders were included in the analysis. The CPM effect during the CS was also analyzed by Wilcoxon signed-rank test.

Inter-individual coefficient of variation (inter-CV) was calculated as standard deviation/mean × 100 (%). Intra-individual coefficient of variation (intra-CV) was calculated as CV=[(∑d2)/2n]/χ×100(%), in which d is the difference between two results obtained from one subject, n is the number of subjects, and x is the mean of the results obtained from all the subjects [27]. The number of non-responders was analyzed with the Mann-Whitney U-test to get the information which is more appropriate for TS; PPT or PPTol, and VAS1.4 or VAS1.6.

The correlations between the VAS peak pain values and relative PPT changes during the CS obtained from all responders in all sessions (correlation coefficient; R) were calculated by means of a least-squares regression analysis.

All data are presented as mean values and standard errors of mean (SEM). The level of significance was set at P< 0.05.

3.1 Pain ratings of the conditioning stimuli

The two-way ANOVA showed that the VAS peak pain values were significantly dependent on the CS (F = 90.018, P<0.001) but not on the session (F =2.664, P =0.104). The post hoc tests showed that there were significant differences among the three CPM techniques; significantly higher (P < 0.001) in the cold stimulus compared to the tourniquet and head band, and significantly higher (P < 0.001) in the tourniquet compared to the head band (cold stimulation: 8.9 ± 0.2, tourniquet: 7.5 ± 0.1, head band: 6.4 ± 0.1) (cm, mean ± SEM) (Fig. 2).

Fig. 2

The visual analogue scale (VAS) peak pain values (mean ± SEM) from three modalities of conditioning stimuli in each session (session I and II). *P < 0.001 vs other modalities of conditioning stimuli. There were significant differences among the three conditioning stimuli (cold pain > tourniquet > head band).

3.2 Test stimulus

4.1 Methodological considerations

Various methodologies have been utilized to evoke and characterize CPM [10]. Pressure pain threshold (PPT) by hand-held algometer tends to be the most sensitive for the test measurement to detect CPM responses [8,28,29]. Moreover, PPT values are stable within the same individual [30] and have a good inter-examiner reliability [31]. Hence it was used in the present study.

The computer-controlled pressure algometry [24,32] allows estimation of the stimulus-response function which is more difficult to obtain by hand-held algometer. The automated system was used for assessing responses from TA, whereas for the other locations it was more adequate to use the hand-held algometer. Therefore, the CPM effects and the inter- and intra-individual CVs were not compared directly between the two algometers.

There were no significant differences in the number of non-responders between PPT and PPTol recording and VAS1.4 and VAS1.6. These results imply that both of these measurements are useful as TS to evaluate CPM.

Gender differences [28] and age differences [33,34] can affect the CPM. When muscle pain is used as CS, more efficient CPM is induced in men [28]. Research on the gender differences of the CPM effect has yielded inconsistent findings [10]. Moreover, it has been reported that the CPM effects of women differ with menstrual phase, suggesting hormonal influence on the CPM [22,35]. In addition, the menstrual phase may also influence the pain response which is lower pain threshold during the luteal phase compared with the follicular phase by CPP [36]. An age-associated decrement of CPM has been reported [33,34]. As only young men have been tested in the study, the issue of gender and age differences was not addressed. However, in the present study we deliberately excluded older subjects and women as these factors would influence the CPM effects. Overall, we consider that we could evaluate the CPM effects more accurately without the influence of age and hormonal factors.

The sample number of this study was 12 for the masseter and forearm and 24 for TA. In view of the fundamental aspects of CPM, we only included the data from CPM-responders in the analysis. However, it is also important to state the proportion of CPM non-responders. Therefore the number and proportion of CPM non-responders are shown in Tables 2,3,4. We cannot deny the possibility of overestimation of the CPM effects and the low reliability because of the low sample. Therefore, additional investigations would be required before the results can be substantiated.

4.2 Modality and assessment site effects of CPM

The CPP [8], tourniquet pain [25,37] and skull pressure pain [26] were applied in the present study as CS. To our knowledge this is the first study to evaluate the CPM effects comparing three different CS modalities together with several assessment sites. In the research concerning CPM and the risk of developing chronic pain, heat pain was applied as both CS and TS [23]. These differences in pain modalities should be noted.

The effects of CPM are known to differ depending on the magnitude and nature of the CS and the stimulated nerve fibers [10,18,19]. The median magnitude of the CPM effect was 29% [10]. In the present study, the CPP induced the most powerful CPM effects (43.6%) in all the measurements.

It seems that the effect of CPM is more intense with stronger intensity of the CS [3,18,19,20,21,22]. In this study, the target level of CS was different among the three modalities of the CS (6 for CPP and tourniquet and 5 for head band) because the pain intensity of mechanical pressure on the head continues to increase gradually for the duration of the compression [26], and actually the VAS peak pain value of head band was 6.4. However the VAS peak pain value of CPP was significantly higher than for the tourniquet and head band. Moreover, there was a significantly positive correlation between the intensity of the CS and the CPM effect. This difference in the intensity of the CS might affect the magnitude of CPM.

Immersion in cold water produces arterial pressor responses [38]. Duschek et al. [39] demonstrate an inverse relationship between blood pressure levels and pain intensity triggered by CPP. From that point of view, hot water seems to be a more appropriate model for tonic pain compared to CPP [40]. On the other hand, CPP, ischemic pain and mechanical pressure have excellent reliability and validity as well as thermal pain [41]. Thus, we consider that the modalities of CS employed in this study were all adequate, and the bigger increases in the threshold determinations and reductions in the VAS ratings induced by CPP are mainly due to the CPM effects.

The central pain-modulating system can be activated not only by noxious stimuli, but also by psychological factors, such as attention, distraction and the environmental demands (emotional arousal) [42–45]. Both distraction and stress can reduce pain [42] and contribute to pain-evoked hypoalgesia, as painful stimuli are not only painful but also distracting and stressful [43]. Also factors such as task difficulty and emotional arousal are implicated in moderating the interruptive function of pain [44]. The intense cold stimulus might shift the subject’s attentional level and have some influence on the CPM effect.

There were significant differences in the CPM effects among the assessment sites; a significantly higher CPM effect at TA than the masseter or forearm. The differences in the CPM effects among the assessment sites might partly be related to the method of evaluation (hand-held or computer-controlled algometer) but also the size of the muscle could influence the response.

4.3 The inter- and intra-individual CV

The inter-individual variation for CPM is important for the experimental and clinical pain research. However, there is no such information available for the reference. One study [46] reported the test–retest reliability of CPM with intraclass correlations (ICCs) and coefficient of repeatability (CR) using hand-held algometry (mechanical TS to the middle finger and trapezius) and occlusion cuff (CS). The results demonstrated that the CPM effect showed no significant differences across test–retest occasions. However, the cuff resulted in large inter-individual variation between test and retest measurements (CR was 1.69 for finger CPM). This is consistent with our result that in general the cuff has larger inter-individual variation than CPP.

In the present study, CPP was found to evoke smaller inter-individual variation compared to the tourniquet pain and mechanical craniofacial pressure. Concerning the factors which might affect the inter-individual CV for CPM, the individual difference of the sensitivity to painful stimuli could be considered besides gender and age differences. Even though we cannot exclude the possibility of variation due to a sensitization of peripheral nociceptors, it would be neglected because any experimental methods might have the differences derived from the general experimental intervention [47].

Regarding the intra-individual variation for CPM, the smallest threshold variability was obtained from the masseter or the forearm regardless of the CS (35.2–40.1%). And for pain ratings, the smallest variability was obtained with CPP (27.0%). There was a general pattern that the CPP leads to smaller intra-CV.

A previous study [21] mentioned that the CPM effect is relatively free of individual variability. A significantly positive correlation of the CPM effect was noted between the two induced temperatures of 46.5 °C and 12 °C, suggesting that individuals who show greater CPM with heat also show greater CPM with cold. The finding supports that the extent of CPM is related to the individual characteristics of the pain modulation response rather than to the modality of CS.

The possible influential factor on the intra-individual variation for CPM is habituation [48]. Though there might be an influence of that between the first and second session, it would be similar for all measurements. Another possible parameter is the influence of examiner [31]. Although the possible bias introduced by different examiners should be taken into account, it does not influence as this experiment was performed by the same examiner (woman). Moreover, the duration of the TS was kept constantly within the study cohort as the same examiner performed the whole experiment.

Another parameter to be investigated for inter- and intra-individual variation is the factor from PPT testing. Previous studies reported that the test–retest reliability of the PPT technique can be guaranteed [46,49]. It has been shown that the CVs for PPT ranged from 20.1% (right parietal) to 47.8% (right temporal) [49] and are relatively large for all measures when compared with other CV findings (e.g.; 18% after 15 min, 14% after 45 min and 29% after 5 weeks [31,50]), but are of similar magnitude to the findings of PPT intra-individual variation [30]. Brennum et al. [30] report that the inter-individual CV for PPT measured by a hand-held algometer was 28% for women and 33% for men, and the intra-individual CV based on repeated PPT measurements with a 1 week interval was 14%. Pud et al. [10] report that the approximated median magnitude of the CPM effect measured by the suprathreshold test-pain reduction was 29% (ranging from 10% to 55%); the increase in the test-pain thresholds was 25% (ranging from 3% to 100%); and the change in the neurophysiological measures was 28.5% (ranging from 10% to 60%). These figures imply that due to the high variation within the 15 studies which measured the changes by test-pain thresholds, continued use of this method in future studies may be questionable.