Prolonged time of after-sensation after experimental pain stimuli despite efficient conditioned pain modulation in patients with chronic neuropathic pain after traumatic nerve injuries in upper extremity

Prolonged time of after-sensation after experimental pain stimuli despite efficient conditioned pain modulation was observed in patients with neuropathic pain after traumatic nerve injuries


Introduction
Why some patients are moderately affected, having no pain after a traumatic nerve injury, whereas others having a similar nerve lesion, develop debilitating chronic neuropathic pain continues to be a mystery. As yet, there is limited research that can identify factors that differentiate between painful and nonpainful neuropathies after traumatic nerve injury. In a previous study, we have demonstrated that about 50% of the patients having a traumatic nerve injury reported chronic pain after their injury, whereas the other half reported no pain. Major nerve injury, younger age, and less time from surgery were predictors for pain after traumatic nerve injuries in the upper extremities. 26 Unfortunately, earlier studies could not detect a specific pattern using quantitative sensory testing (QST) or structural parameters using skin biopsy that would separate the neuropathy patients into painful and nonpainful individuals. 18,21,23,25,38 Impaired conditioned pain modulation (CPM) is an indication of altered descending inhibitory modulation of pain 4,39 and offers the opportunity to test whether patients with neuropathic pain have signs of increased pain sensitivity or abnormal CPM Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
reactions compared with patients with neuropathy without pain. The use of the CPM paradigm in research is reported to be highly relevant for understanding persistent chronic neuropathic pain. 11,40,41 Clear evidence of CPM deficiency 2 has been demonstrated in central neuropathic pain because of spinal injury 1 and in different types of peripheral neuropathic pain, including diabetic neuropathy and postherpetic neuropathy. [31][32][33] The data are conflicting, because in other studies with peripheral neuropathic pain such as diabetic neuropathy 12 or peripheral traumatic nerve injuries, 7 no evidence of deficient CPM was identified.
Could we find some patient-related factors that characterize the 2 different outcomes, helping us to better understand why this occurs? In this study, we have investigated the role of endogenous pain modulation by comparing the variations of the CPM effect as a possible explanatory factor for the differences between painful and nonpainful neuropathy.
To the best of our knowledge, this study is the first that uses a CPM paradigm to explore potential differences in mechanisms between subjects with painful traumatic neuropathies and painless neuropathies.

Participants and methods
This study was performed at the Multidisciplinary Pain Center, Uppsala University Hospital, Sweden. The study was performed in accordance with the ethical principles for medical research involving human subjects that have their origin in the updated Declaration of Helsinki, and the study was approved by the Regional Ethics Committee (approval no: approval for the study was granted by the Regional Ethics Board in Uppsala. Project identity: ICONSS, Dnr: 2015/265; NCT03174665 for Clinical trial organization). Informed consent was obtained from all participants.

Patient recruitment
The patients in this cross-sectional study were recruited from a larger cohort (n 5 669) study with a definite nerve injury in the upper extremities verified and described by the surgeon at the time of surgery for nerve suture or repair. 26 They had previously answered a questionnaire about cold intolerance, pain intensity, previous medication, and the self-report Leeds Assessment of Neuropathic Signs and Symptoms (S-LANSS questionnaire). 26 The participants with pain and S-LANSS $12 (indicating predominantly neuropathic pain) 5 were recruited for the group with neuropathic pain (group A) and those without pain for the group with painless neuropathy (group B). Eligibility for participants was determined only after completion of a health history questionnaire, interview about pain intensity, and a routine clinical neurological examination which confirmed neuropathic pain or painless neuropathy. The inclusion criteria for the participants were as follows: age $18 years, no acute illness or diseases that might impact laboratory performance. The exclusion criteria were as follows: presence of polyneuropathy, diabetes mellitus, peripheral vascular disease, history of malignant disease, and chronic alcohol consumption. All participants were asked to refrain from any pain medication for at least 12 hours before the experimental session. The confirming sensory impairment on examination of the somatosensory system with pain in the innervation territory of a previous intraoperatively verified injured nerve and strongly indicated the diagnosis "definite neuropathic pain" for all the patients in group A. 10,37 The subjects with sensory impairment in the innervation territory of the injured nerve but without pain were included in group B.

Procedures
All participants were informed about the test program after arrival in the laboratory. Participants attended a single appointment. All sessions followed the same procedure and were performed by the same trained examiner who read from a standardized instruction protocol when performing CPM.

Clinical assessment
The participants completed extensive questionnaires which included sociodemographic data, education level, work status, family and medical history, and time from operation. The body mass index was calculated using the formula weight/height 2 (kg/ m 2 ). Baseline brachial resting blood pressure was examined before the experiment was started. All participants were found to be normotensive (111.4 6 8.5/63.5 6 50.0 mm Hg).

Pain assessment and clinical examination
Participants were asked to rate their mean clinical pain over the past week on an 11 point numeric rating scale (NRS). During clinical examination of the somatosensory system, touch was tested with a camel-hair brush (0.5 Somedic, Sweden), pain with a sharp tooth pick, and cold and warm temperature stimuli with warm (40˚C) and cold (25˚C) rollers (Senselab Rolltemp, Somedic). The contralateral uninjured side served as within-subject control.

Quantitative sensory testing
All the subjects were examined by thermal QST in the painful or nonpainful area of the innervation of injured nerve and also the corresponding area contralateral to the injured extremity. For all thermal tests, a thermode (Modular Sensory Analyser; Somedic Sales AB, Hörby, Sweden) with an area of 12.5 cm 2 was used. The initial temperature was 32˚C, and cut-off temperatures were 10 and 50˚C. Warm detection thresholds (WDTs), cold detection thresholds (CDTs), heat pain thresholds, cold pain thresholds (CPTs), and ability to detect thermal and a stimulus-response-function for pinprick sensitivity (mechanical pain sensitivity [MPS]) were calculated as a mean of the 3 measurements on the injured site and noninjured control site in both groups. The 32˚C baseline temperature decreased or increased at a rate of 1.0˚C/second and was stopped by the subject pressing the response button at the moment that warm or cold sensation was perceived. The mean thermal thresholds of 3 stimuli were calculated. The mechanical pain threshold (MPT), MPS including thresholds for pinprick (mechanical detection thresholds [MDT]), pain summation to repetitive pinprick stimuli (wind-up ratio), and ability to detect thermal and a stimulus-response-function for pinprick sensitivity (MPS) parameters were assessed using pinprick stimuli (MRC Systems GmbH, Heidelberg, Germany). Vibration detection thresholds were also tested. Somatosensory phenotypes were determined according to the standardized protocol of the German Research Network Neuropathic Pain. 35

Conditioned pain modulation
The CPM paradigm ( Fig. 1) involved tourniquet pressure test stimulus (TS) applied to one leg, before and after thermal conditioning stimulus (CS) by immersion the uninjured hand in 4˚C cold water.

Test stimulus
The TS was delivered by a tourniquet applied midcalf on the leg corresponding to the noninjured arm and inflated from 60 mm Hg to 100 mm Hg above the systolic blood pressure until the pain intensity (typically 220-250 mm Hg) reported by the patient was over 50 on a 0 to 100 visual analogue scale (VAS). Nociceptive test stimuli were applied for a duration of 120 seconds before (TS1) and after (TS2) the CS.

Conditioning stimulus
The conditioning stimuli was given by letting subjects immerse their noninjured hand up to the wrist in a cold-water bath at 4˚C cooled by a refrigerated water circulator (Somedic, 2015, Sweden) for maximally 1 minute. The CS was applied after TS1 and ended when the patient withdraw the hand from the cold water bath, or maximally for 1 minute. The water level was set at a height of 7 cm, to keep the stimulated area consistent. Time in the cold-water bath (time CS) and time until the pain intensity decreased and the patients became pain free (time off) after removing the hand from cold water were expressed by the area under the curve (AUC CS , AUC time off ) (Fig. 2).
Immediately after the subjects became pain free after CS, an identical test stimulus (TS2) was repeated. The subject was instructed to rate continuously the pain intensity level of the both TS and CS with an electronically eVAS slider until they became pain free. They could terminate the trial at any time if they could not tolerate the painful pressure (120 seconds) or cold water (60 seconds).

Calculations of conditioned pain modulation
To quantify CPM, the deviations of pain ratings from the set point were continuously recorded and summed over time to produce an AUC value. From the start point of the first TS forward, this dependent variable (AUC) of the VAS response over time was calculated for both TS (AUC 1 , AUC 2 ) and CS (AUC CS ). Thus, CPM was calculated as the difference in AUC of pain rating responses between the last TS after CS and the TS before CS (ΔAUC 5 AUC 2 2 AUC 1 ).
The CPM effect (%CPM) is defined as the percent change of the pain intensity evoked by TS induced before and after CS. The formula usually used is as follows: the CPM effect change was (TS1 pain subtracted from TS2 pain) divided by TS1 pain or [(TS2 pain 2 TS1 pain)/TS1 pain] 3 100. The percentage of CPM (% CPM) 5 ΔAUC 3 100/AUC 1 . The CPM effect varies from pain inhibition to facilitation. Therefore, negative CPM scores indicated pain inhibition and positive CPM scores indicated pain facilitation. Efficient CPM was defined as the ability of the individuals to inhibit at least 29% of pain. 34

Measures
Some measures were completed by the subjects before coming to the study such as S-LANSS (Leeds Assessment of Neuropathic Symptoms and Signs). 5

Quality of life
Quality of life was measured at the start of the study with the 36-Item Short-Form Health Survey (RAND-36) a health survey that laps 8 concepts investigating physical and mental status. 17

Quick disability of the shoulder, arm, and hand
QuickDash is a short, reliable, and valid measure of physical function and symptoms related to upper-limb musculoskeletal disorders by shortening the full, thirty-item (Disabilities of the Arm, Shoulder, and Hand [DASH]) Outcome Measure. 3 The last 3 questionnaires were completed when the participants came to the experiment. numbers and percentages for categorical variables. For continuous variables, the Mann-Whitney U test was used. To compare CPM between subjects with neuropathic pain and painless neuropathy and to compare differences between the subjects in the same group, a 2-way analysis of variance (group and side) was performed. A post-hoc unpaired t test was performed for between group comparisons and a post-hoc paired t test for within-group comparisons. For AUC timeoff and time off, we estimated proportional odds models (cumulative probability models using the link function) including the following variables: group (A vs B), age, gender, VAS MAX c, and duration in the cold-water bath. To assess correlation between pain adaptability and pain modulation, Spearman correlation coefficients were used. Most QST-parameters (CDT, WDT, MDT, CPT, and WPT) were log-transformed to conform to a normal (Gaussian) distribution. Quantitative sensory testing was standardized using z-transformation, with z-values above 0 indicating lower thresholds and gain of function, whereas z-scores below 0 indicate hyposensitivity. 35,36 The level of significance was set at a P-value ,0.05.

Study participants, RAND-36, Quick Disability of the Shoulder, Arm, and Hand, hospital anxiety and depression scale
Of 669 subjects previously screened with S-LANSS after traumatic nerve injury, and treated with nerve suture or nerve repair, 26 146 of them received an invitation, and 131 fulfilled the criteria by participating in all tests. Sixty-nine (30 women) subjects had neuropathic pain [group A, 48 years], and 62 (24 women) subjects had a painless nerve lesion [group B, 49 years]. Both groups included subjects with the same types of nerve injuries localized to digital, radial, median, ulnar nerves, and also multiple nerve lesions (P 5 0.643). All the subjects in group A had S-LANSS score $12, and all reported pain. Group B had S-LANSS ,12. Of the subjects with pain resulting from traumatic nerve injury, 80% had no pain medication despite the presence of rather severe pain (NRS $ 6) in 72% of the patients (50/69). Only 15% of subjects with chronic neuropathic pain considered that the medications they had tried had any effect on their pain ( Table 1).

Sensory abnormalities after standard bedside examination
Gain of function (P , 0.0001) evidenced by dynamic mechanical allodynia (P , 0.0001), cold hyperesthesia (P 5 0.0009), pinprick hyperalgesia (P , 0.0001) (Fig. 3), but also loss of function to warm sensation (P 5 0.03) were the sensory abnormalities seen in subjects with pain in comparison with those with nonpainful neuropathic injuries.

Quantitative sensory testing
When the territory of innervation of the injured nerve in subjects with pain was compared with the contralateral healthy side, there were observed thermal differences for CDT, WDT, CPT, TSL (P , 0.0001), but not for heat pain thresholds (P 5 0.08). Elevated mechanical (MDT; P , 0.001), lowered PPT (P , 0.05) and MPT, vibration detection thresholds (P . 0.05 each) were seen. In the individuals with neuropathy without pain, cold detection thresholds (CDT; P , 0.0001; CPT; P , 0.001), MDT (MPT; P , 0.01, each) were elevated compared with the contralateral side and less ability to detect thermal changes were seen (TSL; P , 0.05). Except for statistically significant differences in CPT (P 5 0.03) and lowered PPT in the pain group (P 5 0.01), no other differences in QST were observed between the subjects with pain and painless neuropathy (Fig. 4).

Conditioned pain modulation and conditioned pain modulation effect
The CPM effect or pain inhibition was present in 57 subjects (82%) in group A and 48 subjects (77%) in group B. A CPM effect. 29% indicating a significant analgesic response during CPM 34 was seen in 28 (40%) of the subjects with neuropathic pain and 24 (39%) subjects with neuropathy without pain. A deficit in inhibitory CPM efficacy or pain facilitation was observed in 12 (17%) subjects in group A and 14 (22%) subjects in group B (Fig. 5). There were no significant differences in AUC1 and AUC2 between groups (P . 0.05).

Time for pain to wean off after conditioning stimulus and the response to conditioning stimulus
In group A, 46 (66%) of the subjects could leave their hands in cold water for the full 60 seconds and in group B, 35 (54%) of subjects. Pain after CS lasted longer in patients with neuropathic pain (42 seconds) in comparison with those without pain (19 seconds) (P , 0.0001). As illustrated in Table 2, there were no statistically significant differences in CPM (P 5 0.19), but significant differences were seen regarding AUC time off (P , 0.0001),  found at bedside examination for the subjects included in the group with neuropathic pain (orange) and for those with painless neuropathy (blue). The loss of function describes subjects with decrease and loss of sensation to touch, pinprick, warm and cold and gain of function represents dynamic mechanical allodynia and hyperesthesia to touch, cold and warm hyperesthesia. Gain of function (P*** , 0.0001) evidenced by dynamic mechanical allodynia, cold hyperesthesia, pinprick hyperalgesia, but also loss of function to warm sensation (P# 5 0.03) were the sensory abnormalities seen in subjects with pain in comparison with those with nonpainful neuropathic injuries.
time off (P , 0.0001), and AUC CS (P , 0.0001) between the neuropathic pain and neuropathy without pain groups. After adjusting for age, gender, VAS MAX c, and duration in bath with a cumulative logit model, significant difference between groups A and B, for time off (P 5 0.04) was seen. The odds ratio of a subject in the group with pain recovering later after CS was 1.8 (95% CI 0.97-3.29) times the odds of a subject with painless neuropathy (Fig. 2). No statistical differences were seen between mean peak VAS (VAS CS1 5 87.4 6 17; VAS CS2 5 84. 3 6 22) during CS between groups (P 5 0.46) and no correlation between time off and peak VAS was seen for all the subjects (P 5 0.73, r 5 20.03), as well for group A (P 5 0.39, r 5 20.9).

Discussion
In contrast to our working hypothesis, we found that neuropathic pain patients had a normal function of some aspects of endogenous pain modulation, as no differences in CPM was observed in comparison with subjects with neuropathy without pain. A new finding in this study was that the neuropathic pain patients needed a longer time to recover from the pain experience induced by the conditioning pain stimuli.

Conditioned pain modulation
The potential mechanisms that produce and exacerbate chronic pain remain relatively unknown. 6 Conditioned pain modulation has gained considerable attention among the potential mechanisms that produce pain, primarily because of reports that reduced CPM is associated with the presence of several chronic pain conditions. 9,13,29,30 There are doubts as to the validity of CPM as a biomarker helping us understanding the mechanisms of clinical pain 9 supported by the fact there is no significant correlation between CPM and clinical manifestations of chronic pain. A large proportion of both subjects with neuropathic pain and those with painless neuropathy experienced less pain during the second application of the TS after the administration of CS. However, in both groups, a small proportion of subjects experienced pain facilitation during CPM. The results clearly demonstrated that the presence of CPM cannot be used as a mechanistic sign in differentiating chronic neuropathic pain from neuropathy without pain in this study group. The results underline the speculation presented in another study published by Granovsky 12 that the patients with neuropathic pain (in that case diabetic neuropathy pain) seemed to have "normalized" CPM with the chronicity of the pain syndrome.
The patients in this cross-sectional study was taken from a larger cohort examining risk factors associated with chronic neuropathic pain after nerve trauma followed by surgical repair. 26 All the patients had a definite traumatic nerve lesion, in detail described by the surgeon, who has seen the injured nerve intraoperatively, even if confirmatory tests for nerve lesion, as required for this diagnosis according to the present definition 10,37 were not performed. After all, the "confirmatory tests," such as neurophysiological tests or imaging, are only indirect, "proxy" methods to detect a nerve lesion. The large proportion of subjects with efficient CPM found in both groups might be explained by the small percentage of subjects with severe anxiety, depression, or subjects taking pain medication, factors known to be related with CPM response. 29,43 Most important is the fact that the nerve injuries present after trauma involved the same type of nerves all located in the upper extremities. This decreased the potential for confounding factors in pain etiology and reduced the risk of bias in the conduct and reporting CPM reliability. There are methodological discrepancies among different laboratories in inducing CPM 34 and ambiguous results. Currently, except for some practical recommendations, 42 there is no description of a standardized CPM. Concerning the reliability of our protocol for the assessment of CPM, it was demonstrated previously that cuff algometry for TS and cold bath for CS, the use of sequential protocol testing are appropriate to be used in CPM assessment. 14,42

Time off and response to conditioning stimulus
It has long been known that changes in noxious stimulus intensity have an effect on painful after sensations, 16 but the response after a noxious cold stimulus has not been characterized. It was known that an analgesic mechanism known to quantify endogenous pain modulation is activated during noxious stimulus offset. 15 Among other methods of measuring endogenous pain modulation, one can examine offset analgesia (OA) 7,44 which refers to disproportionately large decreases in pain ratings evoked by small decreases in stimulus intensity. In patients with neuropathic pain, 30 OA was found to be attenuated. If the OA magnitude represents the descending segment of pain sensation after a decrease of noxious stimulus, the slower speed of ). Data are the population mean zscores. Z-scores were calculated in relation to a population of healthy subjects as determined by Rolke et al. 35,36 The horizontal solid lines indicate the 12 and 22 z-score boundaries. A specific QST test is considered abnormal if the testvalue lies above the upper or below the lower boundary. Significant differences between group A and B injured hands were seen for CPT (P 5 0.03) and PPT (P 5 0.01), but no difference between noninjured hands were found as seen in the second graph. The differences between injured and noninjured hand were presented in text. CDT, cold detection threshold; CPT, cold pain threshold; HPT, heat pain threshold; MDT, mechanical detection threshold; MPS, mechanical pain sensitivity; MPT, mechanical pain threshold; PPT, pressure pain threshold; QST, quantitative sensory testing; TSL, thermal sensory limen; VDT, vibration detection threshold; WDT, warm detection threshold. recovery from maximum pain intensity might also be used as a marker of deterioration of descending modulation. 24 Thus, similar to the CPM and OA paradigm, the recovery time after a conditioning cold stimulus can also be used to quantify endogenous pain modulation. Similarly to evidence for a smaller effect of OA 24,30 in chronic neuropathic pain patients 24 as compared with healthy controls, it seemed that the delayed recovery time after a painful stimulus distinguished subjects with neuropathic pain from nonpain subjects. Both CPM and recovery time indicate endogenous pain modulation, but the different results obtained here probably depended on the different neural mechanisms involved in CPM and recovery time. 7 The mechanisms of different endogenous pain modulation pathways differ with probably more brain-driven pain modulation during recovery after cold CS, because of activation of brain regions such as periaqueductal gray, dorsolateral prefrontal cortex, insula, medulla, pons, and cerebellum. 28,44

Quantitative sensory testing
In comparison with the studies on healthy individuals without nerve injuries, 8,20 a large percent of the study participants in our study, who immersed the contralateral arm in the cold bath could not tolerate a 1 minute duration of CS in 4 C water. Cold intolerance is a relatively common symptom after peripheral nerve injury 19,26 and is interpreted as a hyperexcitability in the pain system at the location of neuropathic pain. 28 Except for a predominantly gain of function at bedside examination and elevated values for CPT and lower PPT in subjects with neuropathic pain, 18 no other differences between subjects with pain and painless neuropathy were observed using the same QST protocol. In a recent study, Müller et al. 27 found that hypersensitivity with cold stimulation and lower pressure pain thresholds suggested alterations in central pain processes and were associated with poor recovery after surgery. The presence of later recovery after cold stimulus and reduced tolerance for the cold pressor test 20 in the contralateral side of nerve injury, indicated changes in central processing of pain which may be important contributors to the development of chronic neuropathic pain. 22

Limitations of the study
First, because this sample of subjects was drawn from the subjects with traumatic nerve lesions of the wrist and hand, and in addition operated with nerve suture in a hand surgical setting, the severity and comorbidity findings may be different from that seen in patients with other causes of chronic neuropathic pain. Quantitative sensory testing and CPM data were compared between subjects with neuropathic pain and neuropathy without pain, but not with an independent healthy control group without nerve injury. However, this fact may also be looked on as a strength. Furthermore, this study was a cross-sectional postsurgery study. On the other hand, the patients in our cohort were identified after their initial nerve trauma, and followed up regularly for several years, and all of them having a well-defined definite nerve lesion. In that perspective, the cohort is perhaps unique.

Conclusion
In contrast to our hypothesis, the neuropathic pain patients had a well-functioning CPM, not differing from that of patients with nerve lesion with no pain. A new finding demonstrated in the neuropathic pain patients was a characteristic pattern of delayed recovery time to baseline in reported pain intensity after a cold CS applied to the non-neuropathic hand. This study extends previous findings of the relationship between neuropathic and neuropathy without pain, hopefully contributing to better understanding of chronic neuropathic pain pathophysiology.
In addition, the findings may potentially provide a new way to study and measure the process of endogenous pain modulation, using cold stimuli in relation to painful "after-sensations."

Disclosures
The authors have no conflicts of interest to declare.