The role of stimulation parameters on the conditioned pain modulation response

1 Introduction

Pain modulation processes have been recognized as a key factor in depicting the characteristics of the individual pain profile. One of the most explored mechanisms underlying the pain modulation system is the ‘diffuse noxious inhibitory control’ (DNIC), which is mediated by activation of the spino-bulbo-spinal loop[1]. The DNIC phenomenon is manifested as a decrease in pain sensation evoked by a ‘test-stimulus’ during or following the application of another noxious stimulus, termed the ‘conditioning-stimulus’. This ‘pain inhibits pain’ phenomenon can be measured psychophysically in humans by the conditioned pain modulation (CPM) paradigm [2]. Less efficient DNIC was reported in various populations of idiopathic pain disorders, such as tempomandibular disorder [3,4,5,6]; irritable bowel syndrome[3,7,8];fibromyalgia[9,10,11,12,13]; and tension-type headache [14]. Furthermore, less efficient DNIC was found to be associated with a higher self-report of pain history among healthy subjects [15]. Less efficient CPM was also found to predict the development of chronic post-surgical pain [16,17] and high response to some analgesics [18]. These findings further support the role of prospective assessment of CPM as a valuable laboratory test supporting assessment.

Along with the growing interest in the potential clinical relevance of CPM, a large variability in thenar eminence the techniques applied to induce the CPM response can be found (see a review by Pud et al. [19]); numerous types of ‘test’ and ‘conditioning’ stimuli are used to evoke CPM, with different psychophysical measures (pain threshold, pain scores of supra-threshold stimulus, pain tolerance, or temporal summation), various stimulation modalities (thermal, mechanical, electrical, or ischemic), and different body sites. Moreover, a large variability is found in the CPM extents [19], probably due to the use of different paradigms in the different laboratories. This study was conducted to address the question of consistency of individuals’ expression of their pain modulation when induced by a variety of test protocols. We, therefore, aimed to characterize the results of CPM responses evoked by different paradigms in one cohort.

2 Methods

3 Results

3.2 Correlations between CPM responses across paradigms

Not with standing our expectation that the use of a single cohort would result in correlations, at least between some of the test paradigms, no significant correlations were found between any of the CPM responses (Table 3). It should be noted that no significant correlations were found even before correcting for multiple analysis. According to power analysis, however, this correlation would have survived the Bonferroni correction had the sample size been increased to about 100.

Table 3

Correlations between the six CPM responses (not corrected for multiple comparisons).

	PPT	HP	PP	tTS	mTS
HPT	ρ = 0.096	ρ = 0.340	ρ = 0.051	ρ = –0.023	ρ = 0.081
	p = 0.607	p = 0.061	p = 0.784	p = 0.901	p = 0.664
PPT		ρ = 0.000	ρ = 0.318	ρ = –0.235	ρ = 0.217
		p = 0.998	p = 0.081	p = 0.187	p = 0.241
HP			ρ = 0.203	ρ = 0.227	ρ = –0.267
			p = 0.272	p = 0.204	p = 0.146
PP				ρ = –0.037	ρ = 0.074
				p = 0.841	p = 0.689
tTS					ρ = 0.054
					p = 0.771

Given that in the raw data of the PPT paradigm all subjects demonstrated pain inhibition, it was used to better illustrate the pattern of consistency and the intra-individual characteristics of the CPM response across the various paradigms. Each CPM response was transformed into z score and were plotted to examine their distribution pattern. This was done by assigning black and white colors, respectively, to the most and the least efficient CPM responses (see Fig. 1 ). It can clearly be seen in Fig. 1 that individuals do not sustain a certain profile of CPM efficiency, but rather exhibit a more efficient response in some modalities and a less efficient response in others.

Fig. 1

Characteristics of the CPM distribution for each paradigm. We assign black and white colors to the most and the least efficient CPM responses, respectively according to the CPM responseinthe PPT paradigm. The rest ofthe CPM responses are coloredingray. For each subjects aninconsistent response pattern of more efficient CPM response in some paradigms and a less efficient in others is observed.

4 Discussion

Despite the growing interest in the clinical relevance of descending inhibition, as tested via CPM, it remains unclear as to which psychophysical test paradigm best reveals the individual profile of pain modulation. The recent reviews by Pud et al. [19] and van Wijk and Veldhuijzen [21] on the various measurements and modalities currently used for this purpose point to the lack of studies employing more than one CPM paradigm in the same cohort. The present study was conducted to expand our knowledge about the response patterns of different CPM paradigms assessed in the same sample of healthy subjects. Since most of the relevant pain patients are older than the standard university student cohorts enrolled for experimental pain studies, and since CPM response decreases with age [22,23], we specifically targeted subjects above age 40.

We tried to decrease the variability by using uniform ‘conditioning-stimuli’ because the effect of the ‘conditioning-stimulus’ on the CPM response has already been demonstrated. Granot et al. [24] showed a positive correlation between the CPM responses induced by two types of ‘conditioning-stimuli’. Two other studies revealed higher CPM responses for the cold pressor test as compared to muscle pain, mechanical pain, or tourniquet pain [25,26]. Furthermore, an animal study showed distinct DNIC responses induced by two different ‘conditioning-stimuli’, namely, electrical and inflammatory. Although both paradigms induced robust DNIC responses, they were distinct in their underlying neurotransmitter-related mechanisms. Specifically, the DNIC induced by the electrical stimulation was blocked by a2-adrenoreceptor antagonist, whereas the DNIC induced by hypertonic saline was reversed by both naloxone and a2-adrenoreceptor antagonist [27].

The findings of these studies indicate that the application of different ‘conditioning-stimuli’ may activate different descending inhibitory pathways influencing the extent of the CPM response. Therefore, we focused on the ‘test-stimuli’ by using a variety of modalities and psychophysical measures and maintaining uniform ‘conditioning-stimuli’. Our findings revealed lack of correlations, as well as considerable variability in CPM response, across the various paradigms under these conditions. Thus, these findings point to the central role of the choice of modality and the type of psychophysical measure of the ‘test-stimulus’ in determining the CPM response.

Methodological characteristics can influence the variability in the CPM responses in several ways. First, different stimulation modalities may activate the peripheral nerve fibers (Aβ, Aδ or C) in a different manner, thereby activating different nociceptive pathways that have varying impacts on the central nervous system [28]. Second, a specific modality might evoke a certain cognitive and/or emotional response, based, for example, on the subject’s past pain experience, which may activate a different facet of inhibition [29]. Finally, some of the inter-modality variability may also be attributed to genetic factors. For example, Diatchenko et al. [30] noted that the COMT polymorphism of met/met homozygotes is associated with greater thermal temporal summation as compared to that of val/val homozygotes. Such an association, however, was not observed in response to pressure or ischemic pain modalities. Thus, the capacity of pain modulation is a multifaceted trait that can be expressed differently in response to the application of different CPM paradigms. As such, it is expected that various CPM test paradigms will provide different levels of sensitivity and specificity in characterizing the endogenous modulation capacity of individual patients, though this outcome cannot be discerned by the current study performed with healthy subjects.

Comparative studies of various ‘test-stimuli’ in the CPM paradigm have been minimally done. Lindstedt et al. [31] examined the CPM response assessed by three types of ‘test-stimuli’: pressure pain threshold, heat pain, and nociceptive flexion reflex. A significant correlation was found between the pressure pain threshold and heat pain CPM responses. However, no correlations were found with the nociceptive flexion reflex CPM response and the aforementioned CPM paradigms. Our results are in partial agreement with this report, and this discrepancy may be attributed to the different methodologies used in CPM assessment. In their study, the ‘conditioning-stimulus’, ischemic pain, lasted for 2 min and during this period, the three ‘test-stimuli’ were delivered in a non-randomized order and with only 1-min interstimulus breaks. Moreover, the CPM for the heat pain was calculated only on the basis of pain ratings reported after 30s [31]. These differences in the observed correlations emphasize the effects of methodological considerations on the CPM response. Since our psychophysical data cannot dissect neuroanatomical structures, we can only assume that the CPM response is not a single, uniform response. Therefore, the relevant recommendation for future studies centered on clinical purposes is to use more than one ‘test-stimulus’.

In our concept, there is a continuous spectrum of the CPM response, between pain inhibition on one end, and pain facilitation on the other. This represents the normal variation among individuals. The technicalities of the various testing paradigms determine the range of results per protocol used. Thus, for some protocols the distribution of results is more in the inhibitory range, while for other protocols, the opposite happens. Therefore, we do not think that an overall inhibitory response is a requisite for a certain protocol to be considered relevant. Rather, it is the clinical application that will eventually direct researchers to prefer the use of a certain paradigm. The protocol used in Yarnitsky et al. [16] consisted of a different, more robust, conditioning stimulus that might have contributed to the more inhibitory response to the paradigm.

The significance of our findings should be considered with an awareness of several study limitations. First, we tested a relatively small sample of subjects due to difficulties in recruiting middle-aged, pain-free healthy subjects. Second, although all subjects were assessed in the same setting, the lack of correlations may have been the result of low session-to-session reliability of CPM responses. Third, inter-examiner variation may have affected the CPM responses, given the administration of the CPM paradigms by three different experimenters. However, it should be noted that all three experimenters were trained by the same person and used the same instructions. Moreover, the tests were randomly administered in order to minimize possible inter-test and inter-experimenter bias [32,33]. Despite the a for ementioned limitations, our results highlight important methodological issues in regard to CPM testing.

5 Conclusions

In the current study variability in the CPM responses was observed. This may suggest that the capacity of pain modulation is a multifaceted trait, whose expression varies with the application of different CPM paradigms. Conclusions about which CPM paradigm better reveals the individual efficacyofendogenous analgesia should be reached according to the relevance of a given paradigm in the clinical setting. The CPM paradigm considered as the most relevant in terms of reflecting endogenous analgesia efficacy maybe the one that can predict a patient’s risk of developing a pain disorder or that can provide greater benefit from a specific pain intervention. However, this understanding has not yet been fully achieved. Until more solid clinically relevant data are available, CPM assessment in the clinical setting should be further investigated via the application of several CPM paradigms using different modalities.