Background
Various quantitative sensory testing (QST) modalities have been used to assess particular mechanisms of pain perception in healthy individuals and those with chronic pain. Conditioned pain modulation (CPM) is one QST methodology with purported clinical relevance[
1,
2] and potential to predict the development of chronic pain[
3]. This test paradigm has been previously described as diffuse noxious inhibitory controls but a recent consensus[
4] concluded that the term CPM more accurately describes the testing phenomenon being observed. In short, CPM uses a “pain inhibits pain” test paradigm to activate the descending endogenous analgesia system[
5].
It is believed that one noxious conditioning stimulus inhibits pain from another noxious test stimulus by activating a spinal-supraspinal-spinal loop, resulting in functional inhibitory pain modulation[
1]. The inability of the noxious conditioning stimulus to decrease the pain intensity of the noxious testing stimulus indicates a potential deficiency in the body’s endogenous pain modulatory ability. Several chronic pain populations, including fibromyalgia[
6], tempromandibular joint disorder[
7], and irritable bowel syndrome[
8], have demonstrated decreased CPM efficiency. These results, along with others showing the association of CPM with the development of chronic pain[
3], resolution of pain following treatment[
9], and ability to predict clinical pain[
1], has led researchers to advocate for CPM as a clinically relevant QST modality[
10].
Recently, efforts have gone into standardizing procedures used to assess CPM[
4,
5,
11]; however there is still a need to investigate factors that influence CPM stability. One investigation that examined the test retest reliability of CPM in a small sample of healthy subjects found high reliability among repeated testing within a single session[
11], however the exploration of CPM reliability in different musculoskeletal pain models has not yet been described. This is an important issue which would have implications for use of this measure in clinical populations. Since the evidence shows a dysfunction of endogenous analgesia system among chronic pain groups[
12], it is reasonable to think that CPM reliability could be affected by having the system already engaged with the presence of shoulder pain. Furthermore, since sex differences in pain related responses to experimental pain measures have also been reported[
13,
14], the effect of sex on reliability of CPM also has clinical implications.
Therefore, the present study used two different musculoskeletal pain models to assess whether presence of shoulder pain and gender influence CPM stability within and between sessions. To serve this purpose, this study examined patients with acute and sub-acute shoulder pain preparing to undergo shoulder surgery, and healthy subjects with exercise induced muscle pain (EIMP) at the shoulder[
15,
16]. Understanding how shoulder pain and gender affect pain inhibitory responses, and estimating the error associated with this commonly used measure, will add to our understanding of CPM providing clinical relevance for this particular experimental pain measure.
Discussion
The present study investigated whether gender and pain intensity influence CPM stability in two different musculoskeletal pain models. Our findings suggest that the stability of CPM did not appear to be related to shoulder pain intensity in surgical or exercise induced pain cohorts, however the stability of CPM within and between days differed by sex. Reliability is an essential property of any measurement that needs to be established before the measurement can be used widely in clinical settings. However, psychometric studies of CPM have not been widely reported[
11,
43]. The data reported in this paper are novel because to our knowledge, this is the first study to investigate stability of CPM using the same protocol in subjects with 2 different forms of musculoskeletal pain.
We were interested in determining if the inhibition produced by the conditioning stimulus was reliable within the same session (trial 1 and trial 2) in a cohort of patients before and after shoulder surgery (clinical pain model), and in a healthy cohort before and after pain induction (acute pain model). This is an important issue because a measure with poor reliability is unlikely to be valid[
44]. Using the most common paradigm to evoke endogenous pain inhibition[
4,
5,
45], CPM proved to suppress SHPR as a test stimulus, such that there was a significant inhibition in trial 1 and in trial 2 (within session) in both cohorts. As expected there was less inhibition in trial 2 compared with trial 1 in both cohorts and across both sessions. The starting points in trial 2 (Pre CPM) were lower than trial 1 (Table
1), producing potentially less room for an inhibitory effect (floor effect). This could be explained by lingering inhibition from the first trial, suggesting an inadequate resting period between trials, and corroborating that the CPM inhibitory effect may last longer than 2 minutes, as other authors have previously suggested[
5,
7,
10]. Therefore, future studies should consider a longer rest period if the goal is to have repeated assessment of CPM within the same session.
Stability of CPM was also assessed between sessions. This enabled us to determine if pre and post-surgery changes on pain intensity (for the clinical cohort), and changes on pain intensity during one week (in the healthy cohort), impacted the stability of the CPM paradigm in a controlled situation. While the shoulder pain intensity changed over time (surgical procedure significantly decreased the amount of pain 3 months after the surgery, and EIMP procedure in the healthy cohort induced a significant amount of pain over 5 days), the amount of inhibition remain relatively constant over time, showing that stability of CPM seems not to be affected much by a short period of time and not to be affected by acute or sub-acute changes on pain intensity. The moderate stability of CPM in two musculoskeletal pain models under different conditions of pain intensity, corroborate the assumption that CPM represents a moderately stable response that is largely independent of changes in pain intensity. Even though the purpose of our study was not directly to compare both cohorts (clinical vs healthy), it is interesting to mention that the conditioning stimulus (cold water bath) produced a significant and comparable pain inhibition in both cohorts. These results could suggest that the ongoing pain in the clinical cohort and the induced pain in the healthy cohort did not affect the mechanisms of CPM, reinforcing the idea of the moderately stable nature of this measure, and confirming using cold water bath as a conditioning stimulus as was previously recommended[
4,
5]. In terms of CPM calculations, our results further confirm the importance of reporting the absolute and percent change when reporting CMP[
4]. Not surprisingly, the ICC’s calculated with the percent change had lower reliability values. This could be explained by the different information that each calculation provides, where the absolute difference represents the absolute magnitude of change, and the percent change accounts for the subject’s baseline level of sensitization.
Regarding the influence of sex on CPM stability within and between sessions, the present study revealed that the stability of CPM within and between days differed by sex (Tables
2 and
4). A number of studies using widely different methodologies have investigated sex differences in experimental pain sensitivity[
12] using CPM, or other paradigms. Some researchers suggest less efficient CPM in women than men[
14,
26,
46] while other studies did not detect sex differences in CPM[
47,
48]. Several factors have been proposed to explain sex differences; however the stability of this experimental pain measure across sexes had not been explored before. In this study, ICC’s stratified by sex showed that males from the clinical cohort showed lower scores than females (within session), and the females from the healthy cohort showed lower scores than males (within and between session for the absolute difference). Furthermore, the standard error of measurement (SEM) is directly related to the reliability of a test; that is, the larger the SEM, the lower the reliability of the test and the less precision in the measures taken and scores obtained. Consequently, our results showed a large difference on SEM between males and females, where males from the healthy cohort and females from the clinical cohort had lower levels of standard error of measurement indicating higher levels of score consistency.
The results of the present study in two musculoskeletal pain models suggest that the most stable situation to use a CPM paradigm is in females from the clinical cohort (where changes need to exceed a MDC of 16 to exceed measurement error) and males from the healthy cohort (where changes need to exceed a MDC of 12 to exceed measurement error). This has a direct clinical implication because whether sex affects reliability of CPM, or whether the amount of measurement variation differs across sex may lead to biased clinical results and biased clinical implications. This differences on CPM stability between sex could potentially explain the contradictory results in the literature using CPM as a measure of central pain processing[
3,
9,
12,
49,
50], because a lack of reliability does potentially limit the overall validity of a measure. A speculative explanation for our results could be found on differences on psychological factors associated with this experimental pain measure[
1,
51,
52], differences in expectation between sex[
53], effect of distraction between sex[
26], menstrual cycle effect[
54], however future clinical studies need to explore whether these factors directly affect CPM stability.
Previous studies investigating endogenous pain modulation in chronic pain populations have shown a potential deficiency of pain inhibitory system[
3,
6,
7,
10,
12,
26,
50,
55]. It has also been shown that CPM could be a predictor for post operative pain and potentially sensitive to changes in the central nervous system[
3,
50]. However, since CPM is a proxy measure of central pain inhibitory process, t, it is particularly important to estimate the error associated with this commonly used measure in order to be useful in research and clinical decision making. After performing a reliability analysis in these two cohorts and establishing that CPM is a measure moderately stable independent of changes in pain intensity, we are in a better stage to use this measure in clinical settings (from a reliability stand point). If we think that CPM is a proxy measure of central pain inhibitory system, with a moderate reliability we may speculate that higher stability may be expected when assessing a population with chronic pain (which may have less variability in central sensitization), however future studies need to be performed to test this hypothesis.
Some important limitations of this study will need to be addressed by future research. First, this investigation was part of a larger study; therefore the procedures were not designed solely for the purpose of establishing CPM stability. Second, this investigation would have been enhanced by adding a control condition where CPM testing was repeated on a second pain-free day. In addition, even though the purpose of the study was not to assess the reliability of CPM in the clinical cohort before and after the surgery, our within session analysis could be affected by the effect of drugs. Future studies should consider the potential effect of drugs before and after the surgery. Lastly, some authors have reported significant sex differences in pain report associated with experimenter sex. The present study examined the effect of sex on CPM, however we did not control for experimenter sex effect, menstrual cycle phases or contraceptive use. Future studies should control for these factors, and should investigate the influence of other relevant demographic characteristic (such as race or ethnicity) on CPM stability.
Despite these limitations, the current study represents a novel contribution to the literature by identifying factors that influence CPM stability in two different musculoskeletal shoulder pain models. Evidence suggests that altered central processing of noxious stimuli might be relevant in the pathogenesis of pain disorders[
55‐
58]. However, establishing the stability of a measurement is essential before consider CPM a clinically useful measure. Our results suggest that 1) CPM stability is not related to changes in pain intensity, and 2) there are sex differences for CPM stability within and between days. Applying these results to clinical psychophysics, our study suggest that the fluctuation of shoulder pain intensity did not significantly affect CPM stability, this could reiterate the moderately stable nature of this construct and the relative consistency of this commonly used experimental pain measure.
Competing interest
The authors declare that they have no competing interest.
Authors’ contribution
CV, contributed to conception, design, data analyses, data interpretation, and drafting the article. LK, contributed to conception, design, and drafting the article. RF, contributed to data interpretation and critically revising it for intellectual content. SG procured the funding, contributed to conception, design, analyses and interpretation of data. All authors critically revised the manuscript and discussed the results and commented on it. All authors read and approved the final manuscript.