Assessment of CPM reliability: quantification of the within-subject reliability of 10 different protocols
-
Henrik Bjarke Vaegter
,
Kristian Kjær Petersen
Abstract
Background and aims
Conditioned Pain Modulation (CPM) is a well-established phenomenon and several protocols have shown acceptable between-subject reliability [based on intraclass correlation coefficient (ICC) values] in pain-free controls. Recently, it was recommended that future CPM test-retest reliability studies should explicitly report CPM reliability based on CPM responders and non-responders (within-subject reliability) based on measurement error of the test stimulus. Identification of reliable CPM paradigms based on responders and non-responders may be a step towards using CPM as a mechanistic marker in diagnosis and individualized pain management regimes. The primary aim of this paper is to investigate the frequency of CPM responders/non-responders, and to quantify the agreements in the classification of responders/non-responders between 2 different days for 10 different CPM protocols.
Methods
Data from a previous study investigating reliability of CPM protocols in healthy subjects was used. In 26 healthy men, the test-stimuli used on both days were: Pain thresholds to electrical stimulation, heat stimulation, manual algometry, and computer-controlled cuff algometry as well as pain tolerance to cuff algometry. Two different conditioning stimuli (CS; cold water immersion and a computer-controlled tourniquet) were used in a randomized and counterbalanced order in both sessions. CPM responders were defined as a larger increase in the test stimulus response during the CS than the standard error of measurement (SEM) for the test-stimuli between repeated baseline tests without CS.
Results
Frequency of responders and non-responders showed large variations across protocols. Across the studied CPM protocols, a large proportion (from 11.5 to 73.1%) of subjects was classified as CPM non-responders when the test stimuli standard error of measurements (SEM) was considered as classifier. The combination of manual pressure algometry and cold water immersion induced a CPM effect in most participants on both days (n=16). However, agreement in the classification of CPM responders versus non-responders between days was only significant when assessed with computer-controlled pressure pain threshold as test-stimulus and tourniquet cuff as CS (κ=0.36 [95% CI, 0.04–0.68], p=0.037).
Conclusions and implications
Agreements in classification of CPM responders/non-responders using SEM as classifier between days were generally poor suggesting considerable intra-individual variation in CPM. The most reliable paradigm was computer-controlled pressure pain threshold as test-stimulus and tourniquet cuff as conditioning stimulus. However while this CPM protocol had the greatest degree of agreement of classification of CPM responders and non-responders across days, this protocol also failed to induce a CPM response in more than half of the sample. In contrast, the commonly used combination of manual pressure algometry and cold water immersion induced a CPM effect in most participants however it was inconsistent in doing so. Further exploration of the two paradigms and classification of responders and non-responders in a larger heterogeneous sample also including women would further inform the clinical usefulness of these CPM protocols. Future research in this area may be an important step towards using CPM as a mechanistic marker in diagnosis and in developing individualized pain management regimes.
1 Introduction
In chronic pain patients, several studies have demonstrated impaired conditioned pain modulation (CPM) compared with pain-free controls [1]. Moreover, the CPM response has demonstrated some promise in predicting future pain status [2], the efficacy of analgesics [3], and non-pharmacological treatment [4] contributing to clinical pain profile and management. Assessment of CPM is frequently demonstrated as a change in a test stimulus response (e.g. increase in pressure pain thresholds or decrease in heat pain ratings) by a painful conditioning stimulus (e.g. cold water) applied contralaterally to the test stimulus [5], although various modes of test and conditioning stimuli have been used. In general, CPM is a well-established phenomenon and several protocols have shown acceptable between-subject reliability (based on intraclass correlation coefficient (ICC) values) in pain-free controls [6, 7]. ICC values is a measure of how much of the total variability (within- and between-subject) is explained by between-subject variability, thus reflecting the protocols’ ability to differentiate CPM responses between subjects.
Despite a massive increase in the number of studies investigating the magnitude and between-subject reliability of different CPM paradigms in humans, the authors of a recent systematic review investigating test-retest reliability of CPM [6] concluded that the vast majority of studies did not consider measurement error of the test stimulus (e.g. standard error of measurement, SEM) as some of the change in test stimulus during the conditioning stimulus may be due to measurement error (e.g. habituation). Thus, it was recommended that future CPM test-retest reliability studies should explicitly report CPM reliability based on responders and non-responders (within-subject reliability) based on measurement error of the test stimulus.
A previous experimental cross-over study investigated the between-days reliability of eight different CPM protocols in 26 healthy subjects and reported poor to fair between-subject test-retest reliability across 2 different days for all protocols (all ICCs <0.54) [5]. Although the coefficient of variations for the different test stimuli used were reported as a measure of intra-individual variation between two repeated test stimuli without conditioning stimulus, no information on frequency of CPM responders and non-responders considering these intra-individual variations were reported. As identification of CPM paradigms with adequate within-subject reliability may be a further step towards individualized pain medicine [8], these data provide a good opportunity to describes the frequency of CPM responders versus non-responders across these 10 different CPM protocols considering the standard error of measurement (SEM) for the test stimuli.
Thus, the primary aim of this explorative analysis was to investigate the frequency of CPM responders/non-responders, and to quantify the agreements in the classification of CPM responders versus non-responders using SEM as classifier between 2 different days for 10 different CPM protocols.
2 Methods
This explorative analysis was performed using data on pain sensitivity assessed at 2 different days with four different test modalities before and during two different conditioning stimuli included in a previous experimental cross-over study [5]. Assessments were performed in 26 healthy men (mean age of 25.3±5.6) years (range: 18–42) who provided written informed consent and the experimental study was conducted in accordance with the Declaration of Helsinki, and approved by the local ethical committee (N-20150055).
2.1 Procedure
A complete description of the methodology has been published previously [5]. In short, subjects participated in 2 identical test days at approximately the same time of the day and separated by 1–3 weeks. On each day, pain thresholds to electrical stimulation (Digitimer DS5; Digitimer, Welwyn Garden City, UK), heat stimulation (ATS thermode; Medoc Advanced Medical Systems, Ramat-Yishai, Israel), manual algometry (Somedic Sales AB, Horby, Sweden), and computer-controlled cuff algometry (Nocitech, Aalborg, Denmark) as well as the pain tolerance to computer-controlled cuff algometry (Nocitech, Aalborg, Denmark) were assessed on the dominant side of the body in randomized and counterbalanced order. Assessment of pain sensitivity to the different modalities was assessed four times (two baseline assessments and two assessments during conditioning pain stimuli) each day, before and during two different conditioning stimuli, respectively. Conditioning pain stimuli were induced by cold water immersion and a computer-controlled tourniquet on the non-dominant side of the body in randomized and counterbalanced order. During the conditioning stimuli, subjects used a 10 cm electronic visual analog scale (VAS) to rate the pain intensity due to the conditioning stimulus. Zero and 10 cm extremes on the VAS were defined as “no pain” and as “maximal pain”, respectively. The conditioning stimuli were separated by 45 min to avoid carry-over effects. Previous studies have consistently demonstrated short-lasting hypoalgesic effects of CPM of less than 15 min [9, 10].
2.2 Classification of CPM responders versus non-responders
Subjects who had an increase in the test stimulus response during conditioning which was larger than the SEM of each of the test stimuli for the four baseline measurements (two in each session) was classified as CPM responders and subjects who did not have an increase in the test stimulus response during conditioning larger than the SEM of the test stimulus was classified as CPM non-responders.
2.3 Statistical analyses
Data were analyzed using SPSS Statistics, version 24 (IBM, Armonk, NY, USA).
2.3.1 Baseline pain sensitivity assessments, SEM for the different pain tests and pain intensity during CS
Mean and standard deviations (SD) were calculated for the four baseline pain test measurements, and the absolute difference between two baseline measurements within the same day were investigated with 95% confidence intervals (CI). Repeated measures analysis of variances (RM-ANOVAs) between the four baseline pain test stimuli was performed to investigate potential systematic bias across time. Then, the group SEM for each pain test stimuli were estimated as the square root of the mean square error term in the RM-ANOVA output from the SPSS Statistics [11]. Similarly, mean and SD were calculated for the pain intensity reported during cold water immersion and tourniquet cuff as conditioning stimuli, and the absolute difference between days were investigated with 95% CI. Intraclass correlation coefficients (ICCs) based on a single rating, absolute agreement, 2-way mixed effect model were used reflecting the ability of the different pain tests to differentiate values between individuals. An ICC above 0.75 was taken as excellent reliability, 0.40–0.75 was fair to good reliability, and less than 0.40 defined poor reliability [12].
2.3.2 CPM responders versus non-responders
The numbers of CPM responders versus non-responders at day 1 and day 2 were calculated for the different CPM protocols based on SEM as classifier as described. Cohen’s kappa coefficient was used to investigate agreement between which subjects were classified as CPM responders versus non-responders on the 2 days for each of the different CPM protocols. p-Values less than 0.05 were considered significant. Kappa values of 0.81–1.0 was interpreted as almost perfect agreement, 0.61–0.80 as substantial agreement, 0.41–0.60 as moderate agreement, 0.21–0.40 as fair agreement, and 0.0–0.20 as poor agreement [13]. In addition, and based on the recommendations for future research on CPM [6], the effect of the conditioning stimuli on the test stimuli was calculated as the absolute change (Day 1 minus Day 2) and relative change [(Day 1 minus Day 2) divided by Day 1]*100% in test stimuli during conditioning compared with before conditioning, and SEM for the CPM responses between days were also calculated.
3 Results
3.1 Baseline pain sensitivity assessments, SEM for the different pain tests and pain intensity during CS
Baseline assessments during the 2 days with the five different pain test stimuli (Table 1) showed no systematic errors in day 1 or day 2, as indicated by the 95% CI of the mean differences, where zero lies within the intervals. However, significant differences in heat pain thresholds and pressure pain threshold assessed with cuff algometry across the four measurements were noted. The test-retest reliability of electrical pain threshold, PPT, cPPT and cPTT were excellent with ICCs above 0.80. The test-retest reliability of heat pain threshold was fair to good.
Absolute and relative within-day test-retest reliability for five different pain test stimuli.
| Test stimulus | Day 1 Test 1 Mean+SD |
Day 1 Test 2 Mean+SD |
Absolute difference Day 1 Mean+SD (95% CI) |
Day 2 Test 1 Mean+SD |
Day 2 Test 2 Mean+SD |
Absolute difference Day 2 Mean+SD (95% CI) |
p-Value between all 4 tests | ICC (95% CI) | Standard error of measurement |
|---|---|---|---|---|---|---|---|---|---|
| Electrical pain threshold (mA) | 4.13±2.40 mA | 3.84±2.19 mA | 0.29±0.77 mA (−0.02 to 0.60) | 4.25±2.29 | 4.23±2.22 | 0.02±0.68 (−0.26 to 0.29) | 0.53 | 0.93 (0.88–0.97) | 0.58 mA |
| Heat pain threshold (°C) | 43.33±2.1°C | 43.20±2.36°C | 0.13±1.38°C (−0.43 to 0.68) | 44.03±1.64 | 43.93±1.77 | 0.09±1.34 (−0.45 to 0.64) | 0.008 | 0.70 (0.54–0.84) | 1.03°C |
| PPT (kPa) | 454±145 kPa | 470±138 kPa | −16±89 kPa (−52 to 20) | 495±184 kPa | 480±175 kPa | 15±53 kPa (−6 to 36) | 0.183 | 0.82 (0.71–0.91) | 68 kPa |
| cPPT (kPa) | 23.68±7.87 kPa | 23.37±8.17 kPa | 0.31±5.71 kPa (−1.99 to 2.62) | 24.99±8.85 kPa | 26.83±9.46 kPa | −1.84±5.46 kPa (−4.05 to 0.36) | 0.012 | 0.76 (0.61–0.87) | 4.07 kPa |
| cPTT (kPa) | 51.38±14.07 kPa | 49.42±13.99 kPa | 1.97±8.17 kPa (−1.33 to 5.26) | 51.28±14.72 kPa | 51.65±16.33 kPa | −0.37±5.56 kPa (−2.61 to 1.87) | 0.510 | 0.84 (0.74–0.92) | 5.91 kPa |
-
PPT=pressure pain threshold assessed with manual algometry; cPPT=pressure pain threshold assessed with computer-controlled cuff-algometry; cPTT=pressure pain tolerance assessed with computer-controlled cuff-algometry. p-Value based on repeated-measures ANOVA with the four time points as repeated measure. ICC based on two-way mixed-effects model with absolute agreement, single measures.
Based on the SEM as classifier, the minimal differences in pain thresholds a subject needed to be considered a CPM responder were 0.58 mA for electrical stimulation, 1.03°C for heat, 68 kPa for manual algometry, and 4.07 kPa for computer-controlled cuff algometry, respectively. The minimal differences in pain tolerance a subject needed to be considered a CPM responder was 5.91 kPa for computer-controlled cuff algometry.
For the conditioning stimuli, cold water immersion was rated as significantly more painful on the electronic VAS compared with tourniquet cuff across days and pain test stimuli, but no significant differences between days were found for either of the CS (Table 2). It was noted that the ICC values of pain intensity ratings (VAS scores) between days during tourniquet cuff were somewhat higher compared with cold water immersion.
Absolute and relative between-day test-retest reliability for pain intensity rated during the conditioning stimuli (cold pressor test and tourniquet cuff) in combination with five different pain test stimuli.
| Conditioned stimulus | Test stimulus | Pain intensity (VAS: 0–10 cm) Day 1 Mean+SD |
Pain intensity (VAS: 0–10 cm) Day 2 Mean+SD |
Absolute difference in VAS Mean+SD (95% CI) |
p-Value (between day 1 and day 2) | ICC (95% CI) |
|---|---|---|---|---|---|---|
| Cold pressor test | Electrical | 7.8±1.3a | 7.8±1.4a | 0.001±0.97 (−0.38 to 0.40) | 0.97 | 0.77 (0.54–0.89) |
| Heat | 7.9±1.9a | 8.0±1.5a | −0.13±1.52 (−0.71 to 0.51) | 0.74 | 0.60 (0.28–0.80) | |
| PPT | 7.6±1.6a | 7.6±1.4a | 0.02±1.43 (−0.53 to 0.57) | 0.96 | 0.59 (0.27–0.80) | |
| cPPT | 7.5±1.6a | 7.4±1.8a | 0.12±1.60 (−0.54 to 0.77) | 0.72 | 0.55 (0.20–0.77) | |
| cPTT | 7.5±1.6a | 7.4±1.8a | 0.12±1.60 (−0.54 to 0.77) | 0.72 | 0.55 (0.20–0.77) | |
| Tourniquet cuff | Electrical | 6.0±2.1 | 6.0±2.0 | −0.01±1.60 (−0.66 to 0.64) | 0.97 | 0.71 (0.44–0.86) |
| Heat | 5.8±2.0 | 5.8±2.0 | −0.05±1.48 (−0.65 to 0.54) | 0.85 | 0.74 (0.50–0.88) | |
| PPT | 6.0±2.2 | 6.0±2.1 | −0.05±1.7 (−0.75 to 0.65) | 0.88 | 0.69 (0.41–0.85) | |
| cPPT | 6.0±1.9 | 5.9±2.0 | 0.12±1.22 (−0.37 to 0.61) | 0.62 | 0.81 (0.62–0.91) | |
| cPTT | 6.0±1.9 | 5.9±2.0 | 0.12±1.22 (−0.37 to 0.61) | 0.62 | 0.81 (0.62–0.91) |
-
Pain intensity was rated on a 10 cm electronic visual analog scale (VAS) with 0 and 10 cm extremes on the VAS defined as “no pain” and as “maximal pain”, respectively. PPT=pressure pain threshold assessed with manual algometry; cPPT=pressure pain threshold assessed with computer-controlled cuff-algometry; cPTT=pressure pain tolerance assessed with computer-controlled cuff-algometry; VAS=Visual Analog Scale. (ap<0.001) Significant difference between pain intensity ratings due to cold pressor test and tourniquet cuff.
3.2 CPM responders versus non-responders
3.2.1 Cold water immersion as conditioning stimulus
As illustrated in Table 3, the numbers of CPM responders versus non-responders at day 1 and day 2 for the different pain test stimuli showed large variation in number of subjects classified as CPM responders on both days ranging from five subjects (19.2%) subjects when computer-controlled pressure pain tolerance was used as test stimulus to 16 subjects (61.5%) subjects when manual pressure pain threshold was used as test stimulus. Across the five different test stimuli 11.5% to 46.2% were classified as CPM non-responders. There was no significant agreement in the classification of CPM responders between day 1 and day 2 for any of the test stimuli in combination with cold water immersion (κ<0.31, p>0.1).
Crosstabulations of the CPM responders and non-responders at day 1 and day 2 assessed with cold pressor test in combination with five different pain test stimuli.
| CPM responders day 2 |
||
|---|---|---|
| Yes | No | |
| A. Electrical pain threshold as test stimulus (EPT) | ||
| Cohen’s kappa coefficient: [κ=0.08 (95% CI, −0.31 to 0.46), p=0.70] % agreement (yes or no on both days): 53.8% |
||
| CPM responders day 1 | ||
| Yes | 7 | 6 |
| No | 6 | 7 |
| B. Heat pain threshold as test stimulus (HPT) | ||
| Cohen’s kappa coefficient: [κ=0.31 (95% CI, −0.05 to 0.66), p=0.11] % agreement (yes or no on both days): 65.4% |
||
| CPM responders day 1 | ||
| Yes | 10 | 6 |
| No | 3 | 7 |
| C. Manual pressure pain threshold as test stimulus (PPT) | ||
| Cohen’s kappa coefficient: [κ=0.28 (95% CI, −0.13 to 0.69), p=0.15] % agreement (yes or no on both days): 73.1% |
||
| CPM responders day 1 | ||
| Yes | 16 | 4 |
| No | 3 | 3 |
| D. Computer-controlled pressure pain threshold as test stimulus (cPPT) | ||
| Cohen’s kappa coefficient: [κ=0.09 (95% CI, −0.27 to 0.46), p=0.62] % agreement (yes or no on both days): 53.8% |
||
| CPM responders day 1 | ||
| Yes | 6 | 8 |
| No | 4 | 8 |
| E. Computer-controlled pressure pain tolerance as test stimulus (cPTT) | ||
| Cohen’s kappa coefficient: [κ=0.26 (95% CI, −0.11 to 0.63), p=0.17] % agreement (yes or no on both days): 65.4% |
||
| CPM responders day 1 | ||
| Yes | 5 | 3 |
| No | 6 | 12 |
-
Responders and non-responders are classified based on the standard error of measurement (SEM) for two repetitive test stimulus assessments without conditioning. Responders are defined as an increase in the test stimulus during the conditioning stimulus larger than the test stimulus before the conditioning stimulus plus 1 SEM. PPT=pressure pain threshold assessed with manual algometry; cPPT=pressure pain threshold assessed with computer-controlled cuff-algometry; cPTT=pressure pain tolerance assessed with computer-controlled cuff-algometry.
3.2.2 Tourniquet cuff as conditioning stimulus
The numbers of CPM responders versus non-responders classified at day 1 and day 2 for the different pain test stimuli when tourniquet cuff was used as conditioning stimulus (Table 4) also showed some variation in number of subjects classified as CPM responders on both days ranging from two subjects (0.1%) subjects when electrical or heat pain thresholds or computer-controlled pressure pain tolerance were used as test stimuli to six subjects (23.1%) subjects when manual pressure pain threshold was used as test stimulus. Across the five different test stimuli 30.8% to 73.1% were classified as CPM non-responders. Examination of the agreement in classification of CPM responders versus non-responders between days showed a significant agreement only for CPM assessed with computer-controlled pressure pain threshold as test stimulus [κ=0.36 (95% CI: 0.04 to 0.68), p=0.037], indicating fair agreement in classification of CPM responders versus non-responders with SEM as classifier between 2 different days when assessed with this CPM protocol.
Crosstabulations of the CPM responders and non-responders at day 1 and day 2 assessed with tourniquet cuff in combination with five different pain test stimuli.
| CPM responders day 2 |
||
|---|---|---|
| Yes | No | |
| A. Electrical pain threshold as test stimulus (EPT) | ||
| Cohen’s kappa coefficient: [κ=0.33 (95% CI, −0.13 to 0.79), p=0.09] % agreement (yes or no on both days): 80.8% |
||
| CPM responders day 1 | ||
| Yes | 2 | 2 |
| No | 3 | 19 |
| B. Heat pain threshold as test stimulus (HPT) | ||
| Cohen’s kappa coefficient: [κ=0.235 (95% CI, −0.14 to 0.61), p=0.15] % agreement (yes or no on both days): 73.1% |
||
| CPM responders day 1 | ||
| Yes | 2 | 6 |
| No | 1 | 17 |
| C. Manual pressure pain threshold as test stimulus (PPT) | ||
| Cohen’s kappa coefficient: [κ=0.08 (95% CI, −0.30 to 0.46), p=0.69] % agreement (yes or no on both days): 53.8% |
||
| CPM responders day 1 | ||
| Yes | 6 | 5 |
| No | 7 | 8 |
| D. Computer-controlled pressure pain threshold as test stimulus (cPPT) | ||
| Cohen’s kappa coefficient: [κ=0.358 (95% CI, 0.04 to 0.68), p=0.037] % agreement (yes or no on both days): 69.2% |
||
| CPM responders day 1 | ||
| Yes | 5 | 7 |
| No | 1 | 13 |
| E. Computer-controlled pressure pain tolerance as test stimulus (cPTT) | ||
| Cohen’s kappa coefficient: [κ=0.20 (95% CI, −0.23 to 0.62), p=0.32] % agreement (yes or no on both days): 73.1% |
||
| CPM responders day 1 | ||
| Yes | 2 | 4 |
| No | 3 | 17 |
-
Responders and non-responders are classified based on the standard error of measurement (SEM) for two repetitive test stimulus assessments without conditioning. Responders are defined as an increase in the test stimulus during the conditioning stimulus larger than the test stimulus before the conditioning stimulus plus 1 SEM. PPT=pressure pain threshold assessed with manual algometry; cPPT=pressure pain threshold assessed with computer-controlled cuff-algometry; cPTT=pressure pain tolerance assessed with computer-controlled cuff-algometry.
On a group level, no significant differences were found in magnitude of the CPM responses between days for any of the CPM protocols [Table 5; F(1,25)<3.09, p≥0.09].
Absolute and relative between-day test-retest reliability for CPM assessed with cold pressor test and tourniquet cuff in combination with five different pain test stimuli.
| Conditioned stimulus | Test stimulus | CPM response Day 1 Mean+SD (95% CI) |
CPM response Day 2 Mean+SD (95% CI) |
Absolute difference in CPM response Mean+SD (95% CI) |
Relative difference in CPM response Mean+SD (95% CI) |
p-Value | Standard error of measurement |
|---|---|---|---|---|---|---|---|
| Cold pressor test | Electrical | 0.83±0.99 (0.43–1.23) | 0.73±0.72 (0.44–1.02) | 0.1±1.16 (−0.38 to 0.57) | 57.98±369.00 (−91.06 to 207.03) | 0.68 | 0.82 |
| Heat | 1.55±1.26 (1.04–2.06) | 1.11±0.92 (0.74–1.48) | 0.44±1.48 (−0.15 to 1.04) | 132.92±474.04 (−58.55 to 324.39) | 0.14 | 1.04 | |
| PPT | 153±141 (96–210) | 123±101 (83–164) | 30±124 (−21 to 79) | 13±284 (−102 to 128) | 0.24 | 88 | |
| cPPT | 5.42±7.22 (2.50–8.33) | 4.12±6.53 (1.48–6.75) | 1.30±7.25 (−1.63 to 4.23) | 135.23±246.61 (35.62–234.83) | 0.37 | 5.13 | |
| cPTT | 2.70±8.04 (−0.55 to 5.95) | 5.08±5.91 (2.69–7.46) | −2.38±7.00 (−5.20 to 0.45) | −149.69±774.68 (−462.59 to 163.21) | 0.10 | 4.95 | |
| Tourniquet cuff | Electrical | 0.026±0.84 (−0.31 to 0.37) | 0.25±0.61 (0.001–0.49) | −0.22±0.97 (−0.61 to 0.17) | −51.71±493.87 (−251.19 to 147.77) | 0.26 | 0.69 |
| Heat | 0.41±1.29 (−0.11 to 0.93) | −0.01±1.19 (−0.49 to 0.47) | 0.42±1.27 (−0.09 to 0.93) | 10.37±133.18 (−43.42 to 64.16) | 0.10 | 0.89 | |
| PPT | 76±111 (31–121) | 75±80 (43–108) | 1±135 (−54 to 55) | 118±313 (−8 to 245) | 0.98 | 95 | |
| cPPT | 3.24±5.58 (0.99–5.49) | 1.59±4.35 (−0.19 to 3.31) | 1.68±4.86 (−0.29 to 3.64) | 3.52±254.92 (−99.44 to 106.49) | 0.09 | 3.44 | |
| cPTT | 2.08±5.23 (−0.03 to 4.20) | 1.59±6.41 (−1.0 to 4.18) | 0.49±7.68 (−2.61 to 3.59) | 111.10±465.87 (−77.06 to 299.27) | 0.75 | 5.43 |
-
CPM calculated as absolute difference between test stimuli with and without conditioning stimuli; PPT=pressure pain threshold assessed with manual algometry; cPPT=pressure pain threshold assessed with computer-controlled cuff-algometry; cPTT=pressure pain tolerance assessed with computer-controlled cuff-algometry.
4 Discussion
This study investigated the frequency of CPM responders/non-responders, and quantified the agreements in the classification of responders/non-responders between 2 different days for 10 different CPM protocols. Across the studied CPM protocols, a large and variable proportion (from 11.5–73.1%) of subjects was classified as CPM non-responders when the test stimuli standard error of measurements (SEM) was considered as classifier. The number of subjects who had an increase in the test stimulus response during conditioning which was larger than the SEM of the test stimulus showed large variations between the different test and conditioning stimuli combinations. Agreements in the classification between the 2 days were in general poor across the protocols. Only when CPM was assessed with computer-controlled pressure pain threshold as test stimulus and tourniquet cuff as the conditioning stimulus a significant and fair agreement in the classification of CPM responders and non-responders between days was noted. In addition, this CPM protocol had the lowest intra-individual variation in CPM responses between the 2 days (3.5%).
It is noteworthy, that the commonly used CPM protocol using manual pressure pain threshold in combination with cold water immersion [6], which had the highest percentage of CPM responders on both days (61.5% on both days), and a relatively low intra-individual variation in CPM response between days (13%) showed a poor and non-significant agreement in classification of CPM responders versus non-responders between days. This could be influenced by the small sample size and the small number of CPM non-responders (n=3) in both days limiting interpretation of the results. Moreover, the relatively low ICC values for the pain intensity experienced during the cold water immersion could influence the reliability of the subsequent CPM response.
This explorative analysis is the first to investigate the numbers of CPM responders versus non-responders using test stimulus SEM as classifier between different days for 10 different CPM protocols. In combination with the previously described between-days test-retest reliability ICC values for the CPM protocols under study [5], these results suggest that considerable intra-individual difference in CPM between days exists. The discrepancies in individual CPM responses between days can be caused by adaptation or expectation effects [14]. Moreover, the variability could also be caused by natural physiological variability, variability from the assessment techniques, or variability introduced by the experimenter, although this would potentially be reduced with standardized computer-controlled techniques.
On a group level, no significant differences in magnitude of CPM responses between day 1 and day 2 were found suggesting no systematic group bias in CPM between days.
In patients with chronic pain, several studies have demonstrated impaired CPM compared with asymptomatic controls [1], however the frequency and agreement of CPM responders considering measurement error of the test stimuli between different days in patients with pain is currently unknown.
Strict standardization procedures and reduced bias induced by the person assessing pain sensitivity e.g. via computer-controlled assessment methodology may reduce the intra-individual variation of CPM protocols.
4.1 Implications
These results clearly indicate that choice of protocol used to assess CPM has an influence on the number of subjects who demonstrate hypoalgesia and on how reliable this response is when tested on different days. These findings may suggest that different pain inhibitory systems are activated depending on the combination of test stimuli (e.g. cutaneous of deeper tissue stimulation) and conditioning stimuli which may trigger different mechanisms. Further research are warranted in this area to better understand potential consequences and treatment implications.
5 Conclusion
For most CPM protocols, a large proportion of subjects were classified as CPM non-responders when the test stimuli standard error of measurement was considered as classifier. Considerable intra-individual difference exists between different days, and the agreements in classification of CPM responders versus non-responders were in general poor. The most reliable paradigm was computer-controlled pressure pain threshold as test stimulus and tourniquet cuff as conditioning stimulus. However while computer controlled pressure pain threshold test stimulus and tourniquet cuff had the greatest degree of agreement of classification of CPM responders and non-responders across days, this protocol also failed to induce a CPM response in more than half the sample. In contrast, the combination of PPT and cold water immersion was more effective in inducing a CPM effect however it was inconsistent in doing so. Further exploration of the two paradigms and classification of responders and non-responders in a larger heterogeneous sample also including women would further inform the clinical usefulness of these CPM protocols. Future research in this area may be an important step towards using CPM as a mechanistic marker in diagnosis and in developing individualized pain management regimes.
-
Authors’ statements
-
Research funding: This work was supported by the Innovation Fund Denmark (136-2014-5), The Shionogi Science Program, and TaNeDS Europe.
-
Conflict of interest: Nocitech is partly owned by the Aalborg University. There are no actual or potential conflicts of interest for any of the authors.
-
Informed consent: All participants provided written informed consent.
-
Ethical Approval: The experimental study was conducted in accordance with the Declaration of Helsinki, and approved by the local ethical committee (N-20150055).
References
[1] Lewis GN, Rice DA, McNair PJ. Conditioned pain modulation in populations with chronic pain: a systematic review and meta-analysis. J Pain 2012;13:936–44.10.1016/j.jpain.2012.07.005Search in Google Scholar PubMed
[2] Yarnitsky D, Crispel Y, Eisenberg E, Granovsky Y, Ben-Nun A, Sprecher E, Best LA, Granot M. Prediction of chronic post-operative pain: pre-operative DNIC testing identifies patients at risk. Pain 2008;138:22–8.10.1016/j.pain.2007.10.033Search in Google Scholar PubMed
[3] Yarnitsky D, Granot M, Nahman-Averbuch H, Khamaisi M, Granovsky Y. Conditioned pain modulation predicts duloxetine efficacy in painful diabetic neuropathy. Pain 2012;153:1193–8.10.1016/j.pain.2012.02.021Search in Google Scholar PubMed
[4] Fingleton C, Smart K, Doody C. Exercise-induced hypoalgesia in people with knee osteoarthritis with normal and abnormal conditioned pain modulation. Clin J Pain 2017;33:395–404.10.1097/AJP.0000000000000418Search in Google Scholar PubMed
[5] Imai Y, Petersen KK, Morch CD, Arendt Nielsen L. Comparing test-retest reliability and magnitude of conditioned pain modulation using different combinations of test and conditioning stimuli. Somatosens Mot Res 2016;33:1–9.10.1080/08990220.2016.1229178Search in Google Scholar PubMed
[6] Kennedy DL, Kemp HI, Ridout D, Yarnitsky D, Rice AS. Reliability of conditioned pain modulation: a systematic review. Pain 2016;157:2410–9.10.1097/j.pain.0000000000000689Search in Google Scholar PubMed PubMed Central
[7] Oono Y, Nie H, Matos RL, Wang K, Arendt-Nielsen L. The inter- and intra-individual variance in descending pain modulation evoked by different conditioning stimuli in healthy men. Scand J Pain 2011;2:162–9.10.1016/j.sjpain.2011.05.006Search in Google Scholar PubMed
[8] Yarnitsky D, Granot M, Granovsky Y. Pain modulation profile and pain therapy: between pro- and antinociception. Pain 2014;155:663–5.10.1016/j.pain.2013.11.005Search in Google Scholar PubMed
[9] Vaegter HB, Handberg G, Graven-Nielsen T. Similarities between exercise-induced hypoalgesia and conditioned pain modulation in humans. Pain 2014;155:158–67.10.1016/j.pain.2013.09.023Search in Google Scholar PubMed
[10] Vaegter HB, Handberg G, Jorgensen MN, Kinly A, Graven-Nielsen T. Aerobic exercise and cold pressor test induce hypoalgesia in active and inactive men and women. Pain Med 2015;16:923–33.10.1111/pme.12641Search in Google Scholar PubMed
[11] Weir JP. Quantifying test-retest reliability using the intraclass correlation coefficient and the SEM. J Strength Cond Res 2005;19:231–40.10.1519/00124278-200502000-00038Search in Google Scholar
[12] Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979;86:420–8.10.1037//0033-2909.86.2.420Search in Google Scholar PubMed
[13] Landis JR, Koch GG: The measurement of observer agreement for categorical data. Biometrics 1977;33: 159–74.10.2307/2529310Search in Google Scholar
[14] Goffaux P, Redmond WJ, Rainville P, Marchand S. Descending analgesia – when the spine echoes what the brain expects. Pain 2007;130:137–43.10.1016/j.pain.2006.11.011Search in Google Scholar PubMed
©2018 Scandinavian Association for the Study of Pain. Published by Walter de Gruyter GmbH, Berlin/Boston. All rights reserved.
Articles in the same Issue
- Frontmatter
- Editorial comment
- Support for mirror therapy for phantom and stump pain in landmine-injured patients
- Lifting with straight legs and bent spine is not bad for your back
- Bipolar radiofrequency neurotomy for spinal pain – a promising technique but still some steps to go
- Topical review
- Prevalence, localization, perception and management of pain in dance: an overview
- Clinical pain research
- Pain assessment in native and non-native language: difficulties in reporting the affective dimensions of pain
- Colored body images reveal the perceived intensity and distribution of pain in women with breast cancer treated with adjuvant taxanes: a prospective multi-method study of pain experiences
- Physiotherapy pain curricula in Finland: a faculty survey
- Mirror therapy for phantom limb and stump pain: a randomized controlled clinical trial in landmine amputees in Cambodia
- Pain and alcohol: a comparison of two cohorts of 60 year old women and men: findings from the Good Aging in Skåne study
- Prolonged, widespread, disabling musculoskeletal pain of adolescents among referrals to the Pediatric Rheumatology Outpatient Clinic from the Päijät-Häme Hospital District in southern Finland
- Impact of the economic crisis on pain research: a bibliometric analysis of pain research publications from Ireland, Greece, and Portugal between 1997 and 2017
- Measurement of skin conductance responses to evaluate procedural pain in the perioperative setting
- Original experimental
- An observational study of pain self-management strategies and outcomes: does type of pain, age, or gender, matter?
- Fibromyalgia patients and healthy volunteers express difficulties and variability in rating experimental pain: a qualitative study
- Effect of the market withdrawal of dextropropoxyphene on use of other prescribed analgesics
- Observational study
- Winning or not losing? The impact of non-pain goal focus on attentional bias to learned pain signals
- Gabapentin and NMDA receptor antagonists interacts synergistically to alleviate allodynia in two rat models of neuropathic pain
- Offset analgesia is not affected by cold pressor induced analgesia
- Central and peripheral pain sensitization during an ultra-marathon competition
- Reduced endogenous pain inhibition in adolescent girls with chronic pain
- Evaluation of implicit associations between back posture and safety of bending and lifting in people without pain
- Assessment of CPM reliability: quantification of the within-subject reliability of 10 different protocols
- Cerebrospinal fluid cutaneous fistula after neuraxial anesthesia: an effective treatment approach
- Pain in the hand caused by a previously undescribed mechanism with possible relevance for understanding regional pain
- The response to radiofrequency neurotomy of medial branches including a bipolar system for thoracic facet joints
- Letter to the Editor
- Diagnosis of carpal tunnel syndrome – implications for therapy
- Reply to the Letter to the Editor by Ly-Pen and Andréu
- Letter to the Editor regarding “CT guided neurolytic blockade of the coeliac plexus in patients with advanced and intractably painful pancreatic cancer”
- Reply to comments from Ulf Kongsgaard to our study
Articles in the same Issue
- Frontmatter
- Editorial comment
- Support for mirror therapy for phantom and stump pain in landmine-injured patients
- Lifting with straight legs and bent spine is not bad for your back
- Bipolar radiofrequency neurotomy for spinal pain – a promising technique but still some steps to go
- Topical review
- Prevalence, localization, perception and management of pain in dance: an overview
- Clinical pain research
- Pain assessment in native and non-native language: difficulties in reporting the affective dimensions of pain
- Colored body images reveal the perceived intensity and distribution of pain in women with breast cancer treated with adjuvant taxanes: a prospective multi-method study of pain experiences
- Physiotherapy pain curricula in Finland: a faculty survey
- Mirror therapy for phantom limb and stump pain: a randomized controlled clinical trial in landmine amputees in Cambodia
- Pain and alcohol: a comparison of two cohorts of 60 year old women and men: findings from the Good Aging in Skåne study
- Prolonged, widespread, disabling musculoskeletal pain of adolescents among referrals to the Pediatric Rheumatology Outpatient Clinic from the Päijät-Häme Hospital District in southern Finland
- Impact of the economic crisis on pain research: a bibliometric analysis of pain research publications from Ireland, Greece, and Portugal between 1997 and 2017
- Measurement of skin conductance responses to evaluate procedural pain in the perioperative setting
- Original experimental
- An observational study of pain self-management strategies and outcomes: does type of pain, age, or gender, matter?
- Fibromyalgia patients and healthy volunteers express difficulties and variability in rating experimental pain: a qualitative study
- Effect of the market withdrawal of dextropropoxyphene on use of other prescribed analgesics
- Observational study
- Winning or not losing? The impact of non-pain goal focus on attentional bias to learned pain signals
- Gabapentin and NMDA receptor antagonists interacts synergistically to alleviate allodynia in two rat models of neuropathic pain
- Offset analgesia is not affected by cold pressor induced analgesia
- Central and peripheral pain sensitization during an ultra-marathon competition
- Reduced endogenous pain inhibition in adolescent girls with chronic pain
- Evaluation of implicit associations between back posture and safety of bending and lifting in people without pain
- Assessment of CPM reliability: quantification of the within-subject reliability of 10 different protocols
- Cerebrospinal fluid cutaneous fistula after neuraxial anesthesia: an effective treatment approach
- Pain in the hand caused by a previously undescribed mechanism with possible relevance for understanding regional pain
- The response to radiofrequency neurotomy of medial branches including a bipolar system for thoracic facet joints
- Letter to the Editor
- Diagnosis of carpal tunnel syndrome – implications for therapy
- Reply to the Letter to the Editor by Ly-Pen and Andréu
- Letter to the Editor regarding “CT guided neurolytic blockade of the coeliac plexus in patients with advanced and intractably painful pancreatic cancer”
- Reply to comments from Ulf Kongsgaard to our study
