Whiplash Associated Disorders (WAD): Responses to pharmacological challenges and psychometric tests

2.1.1 Inclusion criteria

The subjects had a well-documented whiplash trauma or a whiplash-like accident with a minimum of six months and maximum of five years before inclusion (i.e., WAD grades II–III). They had a persistent pain in the neck – with or without spread of the pain to the head, shoulder, and arm regions – with a pain intensity of ≥40 mm on a 100 mm VAS. The minimum age was set to 18 years.

2.1.2 Exclusion criteria

All subjects were given a MRI of the cervical spine. If affections of medulla and/or nerve roots corresponding to neurological signs in the periphery were discovered, these subjects were excluded from the study. If the subjects had neuroorthopedic surgery done on the cervical spine, they were excluded. Subjects were excluded if they had a significant chronic pain problem before the trauma. Subjects with generalized pain after the trauma were also excluded. That is, subjects were only included if their pain was mainly localized to the neck with or without spread of the pain to the head, shoulder, or arm regions. Subjects were excluded if they had drug addiction problems, exhibited psychotic behaviour, or were pregnant.

2.1.3 Baseline screening

All subjects answered a comprehensive questionnaire including psychometric instruments relevant for chronic pain. The questionnaire asked the subjects whether they had been using analgesics the previous six months. If they answered “yes”, they were asked for type of analgesic and whether they were using it sporadically or daily (Table 1).

2.1.4 Study context

The present study is a part of a larger project that explores the frequency of cervical zygapophyseal joints as a source for the persistent pain in subjects with chronic WAD. This larger project also tested the efficacy of radiofrequency neurotomy of the innervation of the joints.

2.2.1 Pharmacological challenge

2.2.2 Psychometric instruments

The study used the validated Swedish versions of the instruments listed below.

2.3.1 Pharmacological challenges

The mean of the first two assessments (i.e., before and after insertion of the i.v. cannula but before start of the infusion) was considered the baseline assessment. We classified the subjects as placebo responders, responders, and non responders:

Placebo response criterion: ≥50% VAS decrease (compared to base-line) of the local pain intensity on two consecutive assessment points during the placebo test session.

Response criteria: not a placebo responder and ≥50% VAS decrease (compared to baseline) of the local pain intensity on two consecutive assessment points during test session of active drug.

Non response criteria: none of the above.

Similar classifications were made regarding the assessments of pain unpleasantness and percentage pain relief.

A global response was defined as a response to both morphine and ketamine and a global non response as a response to neither morphine nor ketamine. Placebo responders and subjects with incomplete data for deciding a placebo response were excluded when calculating the proportions for morphine, ketamine, global responders, and global non responders.

The following variables were also determined to compare the three substances:

• Area under curve (i.e., the VAS × time (mm min)) for the rating of local pain intensity and unpleasantness.

• Mean pain intensity and unpleasantness decreases (the difference between baseline and the mean of all other assessment points).

• Mean percentage pain relieve (the mean of all values except the two assessments before start of infusion (i.e., baseline)).

• Efficacy (maximum difference between baseline and a single assessment point) for the rating of local pain intensity, unpleasantness and pain on the categorical scale.

• Efficacy regarding the rating of percentage pain relief (maximum value on a single assessment point).

2.3.2 Psychometric tests

The mean values for the different scales of all the psychometric instruments used were calculated.

2.3.3 Correlations pharmacological challenges and psychometric tests

The responses to the pharmacological challenges were coded into three groups according to the assessments of pain intensity during the test sessions. A positive response was defined as ≥50% VAS decrease (compared to baseline) of pain intensity on two consecutive assessment points:

• Placebo responders – positive responders to the placebo infusion;

• Active responders – not a placebo responder and positive responder to morphine and/or ketamine;

• Non responders – neither a placebo responder nor an active responder.

The mean values of the different scales of psychometric instruments were calculated for the different responder groups and then compared.

2.3.4 Statistical tests

Analyses were made using SPSS for Windows (version 19.0 SPSS Inc. Chicago, Illinois, USA). Multivariate analyses were performed using the SIMCA-P+, version 12.0 (Umetrics Inc.). Results in the text and tables are generally given as mean values ± one standard deviation (SD). Analysis of variance for repeated measures (Friedman’s test) was used followed by two-tailed comparisons (Wilcoxon signed ranks test) to determine which time points differed from the baseline or which drugs differed from the placebo. The Wilcoxon signed ranks test was used for other paired comparisons. Kruskal–Wallis and The Mann–Whitney U test were used for non-paired group comparisons.

When investigating the correlations between the different variables, the Principal component analysis (PCA) and Partial least squares or projection to latent structures (PLS) were applied. Principal component analysis (PCA) using SIMCA-P+ was used to extract and display systemic variation in the data matrix. PCA can be viewed as a multivariate correlation analysis. Variables loading on the same component are correlated and variables with high loadings but with different signs are negatively correlated. A component can be considered as a group of intercorrelated variables. Variables with high absolute loadings and that had a 95% confidence interval not equal to zero were considered significant. Significant variables with high loadings (positive or negative) are more important for the component under consideration than variables with lower absolute loading. A component consists of a vector of numerical values between −1 and 1 (referred to as loadings) and obtained significant components are uncorrelated. Variables that have high loadings (with positive or negative sign) on the same component are intercorrelated. Items with high loadings (ignoring the sign) are considered to be of large or moderate importance for the component under consideration. A cross validation technique was used to identify nontrivial components (p). This method keeps part of the data out from the model development to assess the predictive power of the model and was used to test the significance of the components. The obtained components are, per definition, not correlated and are arranged in decreasing order with respect to explained variation. R² describes the goodness of fit - the fraction of sum of squares of all the variables explained by a principal component.

Partial least squares or projection to latent structures (PLS) was used for the multivariate regression analysis of group membership - placebo responders, active responders, and non responders (i.e., PLS-discriminant analysis; PLS-DA) - using the psychological instruments and pain-related variables as regressors. The VIP variable (variable influence on projection) indicates the relevance of each X-variable pooled over all dimensions and Y-variables – the group of variables that best explain Y. VIP ≥0.8 was considered significant. Coefficients (PLS scaled and centred regression coefficients) were used to note the direction of the relationship (positive or negative). Multiple linear regression (MLR) could have been an alternative when regressing group membership, but it assumes that the regressor (X) variables are independent. If such multi-colinearity occurs among the X-variables, the regression coefficients become unstable and their interpretability breaks down. MLR also assumes that a high subject-to-variables ratio is present (5–10). Such requirements are not required for PCA or PLS; in fact, PLS can handle ratios lower than 1.0. In contrast to MLR, PLS also can handle several Y-variables simultaneously.

Outliers were identified using the two powerful methods available in SIMCA-P+: (1) score plots in combination with Hotelling’s T² (identifies strong outliers) and (2) distance to model in X-space (identifies moderate outliers).

A p-value ≤ 0.05 was considered to be statistically significant in all tests.

3.1.1 Drop-outs

One subject withdrew from the study (the informed consent) before the tests were carried out and is not included in the analysis.

3.1.2 Proportions of placebo responders, responders, and non responders

3.1.3 Assessment of pain intensity, unpleasantness, and per cent pain relief over time for the different drugs

There were no differences in the means of the baseline assessments (i.e., before the drug administration) of pain intensity and unpleasantness between the three sessions (Table 3). The non responders had a significantly higher baseline assessment regarding pain unpleasantness (p = 0.047) than the other responder groups. A similar tendency (non-significant; p = 0.080) was noted for pain intensity.

All three substances significantly reduced pain intensity and pain unpleasantness compared to baseline on all assessment points during the test sessions (data not shown). In the initial parts of the test sessions, ketamine reduced pain intensity significantly more than both the placebo and morphine. Morphine was more effective than the placebo on most of the assessment points and in the later parts of the test sessions morphine was more effective than ketamine. A similar pattern was noted for the assessments of percentage pain relief.

Morphine was significantly more effective than the placebo regarding VAS-area under curve for pain intensity and unpleasantness and regarding mean decrease in pain intensity and in mean per cent pain relief. Morphine also showed a tendency (p = 0.055) to be better than the placebo regarding mean decrease in pain unpleasantness (Table 4).

Ketamine was significantly more effective than the placebo for mean decrease in pain unpleasantness and in mean per cent pain relief. For the VAS-area under curve and mean decrease in pain intensity, ketamine showed no difference compared to the placebo (Table 4).

Morphine showed significance versus ketamine for VAS-area under curve for pain intensity (p = 0.018) and a tendency to significance for VAS-area under curve for pain unpleasantness (p = 0.062). There was also a tendency for morphine to be more effective than ketamine regarding mean VAS decrease for pain intensity (p = 0.068).

3.1.4 Efficacy

Morphine and ketamine showed significant differences in efficacy when compared to placebo except for morphine’s ability to decrease pain unpleasantness (Table 5). Ketamine was significantly more effective than morphine regarding the efficacy for pain unpleasantness (p = 0.049) and per cent pain relief (p = 0.037).

3.1.5 Side-effects

Sedation/tiredness (all three drugs, most prominent for midazolam), dizziness (morphine and ketamine), nausea/vomiting (morphine), and paresthesias/numbness (ketamine) dominated among reported side-effects (Table 6). In two subjects who received morphine and in four subjects who received ketamine, the infusions were discontinued because of side effects.

3.1.6 Influence of pre-study use of analgesics

Some data were missing (13%, Table 1) regarding the pre-study use of analgesics.

3.1.7 Influence of the sequence of infusions

We compared the groups where the placebo infusions were given in the first session (n = 32), second session (n = 30), and third session (n = 31). The analysis showed no significant differences with respect to gender, age, duration of symptoms, and baseline assessments of pain intensity and unpleasantness (data not shown). VAS area under curve, mean VAS decrease, efficacy, and mean and maximum per cent pain relief for all three substances did not significantly depend on where the placebo infusions were placed in the sequence. There were no significant impacts of the placebo sequence on the proportions of the responder groups (data not shown).

Similar analyses were made for morphine and ketamine where we grouped the whole sample according to where the drugs were placed in the sequence of infusions. As with the placebo infusion, morphine and ketamine did not significantly influence the main results of the study (data not shown).

4.2.1 Different purposes of intravenous pharmacological tests

Our study displays heterogeneity in a group of subjects with chronic WAD. With the method and definitions used, we are able to draw some conclusions about the heterogeneity, but we can neither make any predictions on the outcome of long-term therapy with the drugs tested nor conclusions regarding specific pain processing mechanisms.

Since the early 1990s, the literature has seen an expansion of reports concerning intravenous (i.v.) pharmacological tests including a multitude of drugs and different pain states. The purposes of the tests differ. Some studies try to elucidate whether results from an i.v. drug test can predict the outcome of a long-term drug therapy [23,38,39,40,41,42]. Other studies use pharmacological tests to predict the outcome of surgical or other invasive procedures [43,44]. A third focus is to use i.v. drug tests, presuming known targets of the drugs, to analyse pain pathophysiology in experimental pain [24,45,46,47,48] as well as in clinical pain states [25,49,50,51,52] and in combinations (induced experimental pain in subjects with a chronic pain state) [11,26,53].

A systematic review found weak or no evidence for the utility of i.v. infusion tests [54]. However, this review focused on evidence for the tests ability to predict the outcome of a long-term drug therapy. The shortcomings of these studies were related to more than just methodological fiaws. For instance, some drugs used for i.v. tests do not have an analogous available drug for long-term therapy or there might be side-effects not seen in the i.v. test but seen after some time during the long-term treatment.

4.2.2 Cutoff points

One aspect for consideration is where to set the cutoff point for the dichotomy response and non response to a pharmacological challenge. We chose the 50% paradigm, i.e., a response was defined as a ≥50% pain decrease on two consecutive assessments during the test session compared to baseline; anything less was considered a non response. Farrar et al. studied the subject by reanalysing former clinical trials [55,56]. They tried to define the degree of change in absolute pain rating scales (e.g., an 11-point pain intensity rating scale, NRS) that best corresponded to significant clinical improvement estimated by some global assessment (by the patient). They found, on average, that a reduction of approximately two points (on an 11-point NRS) or a reduction of approximately 30% represented a significant clinical difference. However, these studies concerned the cutoff point for significant clinical improvement in long-term treatment of chronic pain. Our study has another target: trying to display different patterns of responses to an i.v. pharmacological challenge in a group with a specified chronic pain state. Cohen et al. designed a cutoff point to a 67% pain relief, which gave the best prediction by an i.v. ketamine test for the outcome of an oral dextromethorphan treatment in neuropathic pain [39]. They used the same cutoff point in similar studies regarding fibromyalgia patients [40] and in opioid-exposed patients with persistent pain [41]. Again, this cutoff point was designed for another purpose than what is relevant for our study. To our knowledge, no similar datadriven studies exist that define the cutoff points for i.v. drug tests to elucidate patterns of responses on a group level. We find it reasonable to believe that the 30% cutoff point might be too weak and the 67% cutoff too strong for our purpose. Hence, we chose the 50% paradigm although it is an arbitrary designation to some extent.

4.2.3 Pain intensity and pain unpleasantness

In our study, there were small differences (on a group level) between the two assessment modalities – pain intensity and pain unpleasantness – both regarding the baseline assessments and the proportions of different responses to the substances. Hence, it can be questioned whether it is possible to distinguish between the intensity/sensory and the unpleasantness/affective factors in the perception of pain. An alternative is that these two factors are separable but strongly correlated. Traditionally, the pain experience has been analysed in the following three ways: sensory/discriminative, affective/motivational, and cognitive/evaluative. Fields argues that there are primary and secondary unpleasantnesses [57]. The first is stimulus bound and hence should be analysed as a sensory/discriminative component of pain. In our understanding, one could, for example, ask whether there is any pain that is not unpleasant. The secondary unpleasantness “[...] is a higher level process to which contextual features contribute powerfully resulting in an emotional experience[...]” [57]. The question can be raised regarding what kind of unpleasantness we measure, especially in a pharmacological short-term drug test. We also used a placebo substance (benzodiazepine) that could have influenced the affective pain component.

4.2.4 The possible influence of pre-study use of analgesics

Unfortunately, data were missing regarding the use of pre-study analgesics. However, we did not find any significant influence on the major results of the study. No subject in the sample reported any pre-study use of strong opioids, which, of course, would have influenced the analysis.

We find it difficult to explain why those reporting no use of NSAID and/or acetaminophen have a significant lower mean VAS decrease during the morphine test (p = 0.044). Without any rationale we interpret this as a random result.

Those subjects reporting a daily pre-study use of weak opioids had a significantly higher baseline assessment of pain intensity (p = 0.028). As the subjects were instructed not to take any analgesics for at least 8 h before the test sessions, this result might refiect the absence of their daily used analgesic at the baseline assessment. Together with the short-acting pharmacokinetic profile of ketamine, this could also explain why subjects who used weak opioids on a daily basis had a larger VAS area under curve during the ketamine infusion.

4.2.5 Alternatives in study design

There are some possible alternatives when designing a pharmacological study like the present one. For instance, a study can consider the pharmacokinetic profiles of the drugs and adjust the timetable for assessments of effect accordingly. For the sake of simplicity and for optimizing the blinding procedure, we chose a design with a fixed timetable and fixed points of time for assessment of effect, a strategy that covered most of the elimination half life for the drugs tested. The differences between pharmacokinetic profiles of the drugs were refiected in the results with the short acting ketamine proving more effective in reducing pain in the beginning of the test sessions and morphine proving more effective in the later parts. Morphine was also more prominent in reducing the VAS area under curve, whereas ketamine seemed to be more prominent in efficacy parameters.

4.3.1 The placebo effect per se

The placebo effect is indeed a complex phenomenon. Recent years have seen an expansion of research into the mechanisms of the placebo effect, especially regarding pain (experimental and clinical). One of the main results of this research is the knowledge that a placebo response is a real psychobiological phenomenon where the central nervous system is involved not only on the psychological level but also on a physiological level. It represents a link between a complex mental activity and the body [58]. There is not one specific mechanism responsible for the placebo effect and there is not a single placebo effect, but many. There are different mechanisms for different medical conditions and interventions [58]. When it comes to pain, data indicate that the placebo effect works in part through the opioid-related endogenous pain modulatory descending circuits and in part through the dopamine-related circuits for reward [59,60]. The placebo effect regarding pain is robust and strong. In one study, the placebo was as effective as a hidden i.v. injection of 8 mg morphine [61]. Several studies have found the magnitude of the placebo analgesic effect to be around 2 out of 10 on a visual analogue scale. When viewing the placebo responders selectively, the effect is even more impressive: 3.3–5 out of 10 [60]. Some data indicate that the placebo effect working through the endogenous opioid system also has the capability to be selectively directed to local parts of the body; i.e., it works under a somatotopical structure [62]. Other data indicate that the placebo might work on a spinal as well as a supra-spinal level with impact on the mechanisms we think are responsible for central sensitization; i.e., the placebo effect might mimic the effect we assign to ketamine [63].

This knowledge about the placebo effect raises the following question: How should we interpret the placebo response in a pharmacological drug test? In addition to an active placebo (midazolam), we tested morphine and ketamine and the placebo effect might mimic them both. However, we find it reasonable to exclude the placebo responders when analysing the proportions of the other responder groups, but bearing in mind that among the placebo responders there might be some with “true” morphine-responsiveness, some with “true” ketamine-responsiveness, and some with a combination of the two. In the present study, only two out of 24 placebo responders (pain intensity) selectively responded to the placebo substance.

4.3.2 Midazolam

The present study obtained a higher proportion of placebo responders (26%, pain intensity) compared to an earlier study of WAD-subjects, which had a proportion of approximately 10% when saline was used as a placebo [11]. On the other hand, the frequency of side-effects, especially sedation and tiredness, indicates that our placebo infusions to some extent mimicked the infusions of the other substances, which was why we used midazolam as a placebo substance. The frequency of sedation and tiredness was more prominent for midazolam infusion than for morphine and ketamine (Table 6), indicating that the dosage of midazolam was too high. This finding suggests that future studies may want to consider reducing the dosage.

A possible confounding factor in the analysis of the placebo response could be where the placebo infusions were placed in the sequence – first, second, or third infusion. However, we found that where the infusions were administered in the sequence had no major impact.

We used midazolam as an active placebo substance assuming that it has no inherent analgesic activity. However, the absence of analgesic effects of benzodiazepines (BZ) is not a straightforward matter. It is fairly well documented that BZ has analgesic effects when administered intrathecally [64,65], whereas the effects with systemic administration have been more controversial. Several clinical reports indicate analgesic effects in chronic pain syndromes such as cancer pain, phantom limb pain, and myofascial pain [66]. Knabl et al. [67] speculate that the controversy regarding systemic BZ might be due to studies not distinguishing between acute pain and chronic pain states with hyperalgesia. They state that intrathecal BZ can have antihyperalgesic effect in the absence of any antinociceptive effect on acute pain [68].

4.4.1 Morphine

Morphine is the prototypical µ-receptor agonist presumed to exert its main analgesic effect through a blockade or inhibition/modulation of an on-going nociceptive input to the pain transmission system [69]. Clinically, it is well known that there is a high degree of inter-individual variation of sensitivity for the analgesic effect as well as for side-effects. This variation might be due to genetic variations in the expression and distribution of receptor subtypes [70].

In the present study, 32 of 69 (47%, pain intensity, placebo responders excluded) responded to the morphine infusion according to the assessments of pain intensity. This result is similar to what a smaller study of chronic WAD subjects found [11]. We find it reasonable to interpret these results taken together as showing that at least a subgroup of subjects with chronic WAD has opioid sensitive pain.

4.4.2 Ketamine

Ketamine exerts its analgesic effect mainly through a non-competitive blockade of the NMDA-receptor in the central nervous system even though at higher doses it may interact with µ-opioid receptors and suppress sodium channels [71]. The importance of the NMDA-receptor for a nociceptive or neuropathic input to initiate central sensitization is well documented [17,72,73]. It is also well documented that at least a subgroup of patients with chronic WAD shows signs of a central sensitization [20,21,22].

We used a subanaesthetic low-dose ketamine infusion (0.3 mg/kg), which presumably would rule out any significant analgesic effect other than the NMDA-receptor blockade. Of 69 subjects, 28 (41%, pain intensity, placebo responders excluded) responded to ketamine. This result might indicate the presence of a central sensitization in a subgroup of patients with chronic WAD.

4.4.3 Global responders

Of 69 subjects, 17 (25%, pain intensity, placebo responders excluded) responded to both morphine and ketamine.

Some studies indicate the importance of a peripheral nociceptive input for initiating and maintaining a central sensitization in chronic pain [74,75,76]. It is reasonable to believe that there are variations in the degree of on-going nociceptive input as well as in the degree of central hypersensitivity among subjects with pain such as chronic WAD. In the present study some individuals responded primarily to morphine, some to ketamine, and some to both drugs.

4.4.4 Non responders

An intriguing result of our pharmacological tests is the relatively high frequency of non responders (25 out of 69, 37%, placebo responders excluded). This finding is in line with a smaller study of subjects with chronic WAD that showed a frequency of global non responders of 33% [11]. These figures for global non responders concerning chronic WAD are higher than in comparable studies regarding other pain states: one study on patients with fibromyalgia syndrome showed a frequency of 17% [25] and one study on low-back pain showed 25% global non responders [43].

Hence, one of four subjects in a relatively large group of subjects with chronic WAD did not reach 50% pain relief when challenged with an active placebo infusion, an infusion of morphine (0.3 mg/kg), or an infusion of ketamine (0.3 mg/kg), otherwise documented to have a potency for a high analgesic effect. So far we can only speculate about this result:

• We tested three substances. There might be other targets in the pain processing system not reached by our drugs but relevant, e.g., for central hyperexcitability (sensitization).

• A subgroup of patients with chronic WAD might be extremely biased to the psychosocial side of the biopsychosocial pain model, making these patients out of reach for any significant pain relief provided by analgesic drugs.

• Assessment considerations must be addressed. We placed subjects with a chronic pain state in a rather odd experimental situation, compared to daily life, and asked them for a minute-by-minute estimation of pain intensity, unpleasantness, and per cent pain relief. All this was done during a drug infusion with side-effects probably obscuring the capability of correct estimates to some extent.

The non responders had a significantly higher baseline assessment of their pain unpleasantness compared to the other groups (p = 0.047). There was also a weak trend (p = 0.080) for higher base-line in pain intensity in the non responder group.

In addition to the tendency for the non responder group subjects to rate their baseline pain higher than the other responder groups, the non responders seem to be the worst cases in at least some aspects of the psychometric tests used in this study. This finding could be in line with clinical experience when it comes to treating chronic pain: patients with scores indicating depression, anxiety, maladaptive coping strategies, etc. seem less disposed to respond positively to a single medical intervention such as a drug treatment.

We can only speculate about the reasons why non responders seem to score badly in the psychometric tests. The biopsychosocial model is now widely accepted as a model for understanding chronic pain disorders. The underlying neuromatrix for pain is a complex neurophysiological system mediating all the factors inherent in the model, leading to the final pain experience and behaviour. The ascending somatosensory input from the periphery is modulated in the matrix and descending pathways from the matrix have the potential to facilitate or inhibit the peripheral input at different levels of the central nervous system. The complexity increases when considering the plasticity of the neuromatrix, i.e., an on-going nociceptive or neuropathic input in the system tends to change the matrix in different ways. Furthermore, we have the question of the reversibility of these plastic changes. Finally, we can consider possible genetic variations in how the matrix performs the pain processing mechanisms. In the context of this complexity we have to consider the i.v. infusion of a pain-reducing drug as a rather coarse intervention. It is understandable that some individuals with a chronic pain disorder have other dominating factors in the matrix than those targeted by the drug, leaving the pain experience more or less unaffected by the drug intervention. For some individuals, some of these other factors might be refiected in the psychometric tests.

4.5.1 Sample bias

The sample in the study was recruited from patients referred to a second or third level of health care institutions for management of chronic pain. Hence, those managed in primary care were excluded. This could put into question how representative the sample is for the whole population of chronic WAD. The estimate of the population itself is poorly defined in the literature: different studies show different results regarding incidence and prevalence of chronic WAD as well as regarding the recovery rate from acute WAD [2,3,77]. This context of what is considered to be chronic WAD adds further uncertainty regarding the relevance of a single sample.

4.5.2 Inclusion/exclusion criteria

Some chronic WAD patients develop a generalized musculoskeletal disorder including spontaneous pain in all four quadrants of the body. Studies trying to reveal a relationship between chronic WAD and fibromyalgia show contradictory results [78,79,80]. We decided to exclude from the study those with generalized pain and restricted the sample to those with localized pain in the neck with possible referred pain to the head and/or upper extremities. The reason for this exclusion criteria was that we planned to evaluate the effect of diagnostic blocks of zygapophyseal joints and the efficacy of radiofrequency neurotomy of the innervation of these joints. The results of this will be considered in future papers. However, this exclusion criteria might imply a further restriction in the relevance of the sample in relation to the population of chronic WAD.

4.5.3 Pain mechanisms

In recent years, pain research has focused on genetic variability. Genetic variations are considered to have an impact on the disposition to develop a chronic pain state and how the neuromatrix responsible for processing a nociceptive or neuropathic input is working. Genetic variability is also believed to be of importance when it comes to responses to medical interventions such as analgesic drugs. Hence, this variability is a factor to consider when trying to understand precise pain mechanisms and their treatment in the individual patient as well as in defined groups of chronic pain states. In an ideal world, we know these precise pain mechanisms when meeting the individual patient or when dealing with a group of patients with a defined chronic pain state. We can target this known mechanism in our treatment and restrict our samples for clinical research to those with pain mechanisms apt to respond to the intervention we are investigating. Today, in the clinic as well as in clinical research, we seldom have this precise knowledge of pain mechanisms. Most often we deal with syndromes with poorly defined mechanisms for chronic pain: failed back surgery, low back pain, general myofascial pain, fibromyalgia, temporo-mandibular disorders, CRPS, etc. The underlying variability in pain syndromes leads to poor treatment results. Some impact of the variability can be reduced in clinical research by expanding sample sizes, but treatments are sometimes probably ruled out simply because of poorly defined pain mechanisms in the sample.

Chronic WAD is a pain syndrome with poorly defined pain mechanisms, presumably involving a great deal of variability. When challenging patients with an i.v. infusion of morphine, ketamine, and midazolam (active placebo), we do not know whether the variations in responses reflect variations inherent in the concept of chronic WAD or simply reflect genetically determined dispositions. However, we do know that at least some of the subjects with chronic WAD have signs of central sensitization in their pain processing system [18,21,22]. We also know that the NMDA-receptor, the main target for ketamine, is of major importance in the development of central sensitization [16,17]. We do have data that suggests that morphine and ketamine have different targets for their pain reducing effects [46] andwehave data displaying synergistic effects of morphine and ketamine [24,47,48], which also implies that there are different targets for the substances. Hence, the present study does not reveal precise pain mechanisms but does reveal one aspect of the heterogeneity in the population of subjects suffering from chronic WAD.