Summary This thesis focusses on pain-related fear in patients with chronic pain, more specifically on extinction of fear via exposure in vivo treatment in patients with Complex Regional Pain Syndrome Type I. Chapter 1 starts with a general introduction on pain, fear, learning and exposure in vivo treatment. Following a biomedical approach, pain is traditionally defined as a signal of injury or tissue damage, suggesting a one-to-one relationship between the extent of injury, and the intensity of pain. It has now been demonstrated that the experience of pain is much more complex: on the one hand tissue damage is not always painful (e.g. MRI techniques confirmed that abnormalities are quite common in pain-free individuals), and on the other hand, pain can occur without injury (in most cases of low-back pain, no pathology can be detected). Pain can be better understood as a signal that encourages behavior aimed at escaping a dangerous situation than as a sign of tissue damage. The experience of pain is also modulated by central psychological mechanisms as cognitions and emotions. For example, a constant physiological pain stimulus can be experienced as more or less painful, depending on the instruction that this stimulus is safe or harmful. The interaction of physiological and psychological factors shows that a biopsychosocial perspective is needed to understand the complexity of pain. From that perspective, pain is currently defined as "an unpleasant sensory and emotional experience associated with actual or potential tissue damage or described in terms of such damage". In addition, the complexity of pain increases when over the course of time, the presumed relationship between tissue damage and pain becomes more doubtful. When pain "exists beyond the expected period for healing" it is defined as chronic pain. Paradoxically, protective behaviors, which were adaptive in an acute pain context where tissue damage occurred, can exacerbate the problem once the pain is chronic. The fear-avoidance model states that the reaction of an individual to pain might either be confrontation or avoidance. Most people will evaluate pain as unpleasant, but not as a threat or catastrophe. That is why they proactively confront their pain, resume daily activities and report a reduction in perceived pain in the long run. However, there are also individuals who attribute a catastrophic meaning to the pain experience, for example as a sign of injury or a predictor of aggravating pain. Such catastrophic interpretations give rise to pain-related fear (e.g. fear of pain, fear of movement, fear of (re-)injury). Fear then causes hypervigilance, an elevated state of sensory sensitivity, accompanied by behavior aimed at detecting and avoiding threat. Hypervigilance interferes with attention for the performance of relevant activities and might increase the detection of pain and other aversive bodily signals. An anxious person will also be inclined to respond defensively to pain and, for example, show avoidance behavior. In chronic pain patients, avoidance behavior loses its adaptive function as a protective strategy; the injury has healed, despite persistent pain, and pain is typically no longer a sign of actual danger. Avoidance of relevant daily life activities results in disability, has a negative impact on mood and compromises quality of life. These negative effects of avoidance in turn enhance the pain experience, fueling the vicious fear-avoidance circle. Learning theories help to understand why people become fearful of innocuous stimuli. Acquisition, generalization and extinction of fear can be understood from the principles of classical conditioning (learning associations); the influence this subsequently has on behavior can be understood from operant learning principles (behavior in- or decreases when it is reinforced or punished, respectively). Exposure in vivo is based on classical conditioning principles and is the preferred treatment for pathological fear. In pain-related fear, this treatment aims to reduce disability by challenging and correcting erroneous expectancies about pain and activities. Several RCTs have shown that exposure in vivo is effective in reducing disability, pain-related fear and experienced pain in patients with chronic low back pain. Repeated single-case experimental designs showed comparable results not only for chronic low back pain, but also for other chronic pain conditions, such as whiplash associated disorder, complaints of arm neck and shoulder (CANS), and Complex Regional Pain Syndrome Type I (CRPS-I). CRPS-I is an intriguing clinical manifestation of chronic pain with an unknown pathophysiology. It is characterized by severe pain that is disproportionate to the inciting event, localized in a distal part of an extremity and accompanied by abnormal swelling, changes in skin-color and temperature, as well as in sweating patterns that deviate from normal. When CRPS-I becomes a chronic condition, disability is often progressive over time, despite a variety of medical treatments. As in other chronic pain patients experiencing disability, pain-related fear was found to be present in CRPS-I patients as well, and a first repeated single case experimental design on exposure in vivo showed promising results. Next, a number of points concerning pain-related fear and exposure in vivo are mentioned that require further investigation. These points have been translated into the following research questions, to which this dissertation aims to provide an answer: How can we understand the acquisition and extinction of pain-related fear from a learning theory perspective? Does verbal and visual manipulation of threat influence acquisition, extinction and return of fear in pain-free individuals? Is generalization of exposure in vivo treatment facilitated when many different activities are performed or when a few activities are repeated several times? Is exposure in vivo an effective treatment in reducing disability as primary outcome, but also in reducing pain-related fear and pain intensity, as well as in improving quality of life in patients with CRPS- 1 and pain-related fear? Is exposure in vivo cost-effective in CRPS-I patients with pain-related fear, compared to pain-contingent physical therapy? In each subsequent chapter, one of the above-mentioned research questions is addressed. Chapter 2 focusses on the mechanisms underlying acquisition and extinction of pain-related fear. A narrative review of the literature on fear conditioning and extinction specifically applied to pain-related fear is presented. Classical fear conditioning occurs when an emotionally neutral stimulus becomes associated with an inherently aversive experience, thereby acquiring the capability of evoking a fearful response on its own. The inherently aversive experience is the unconditioned stimulus (US, e.g. pain associated with harm or (re-)injury). The initially neutral stimulus that becomes associated with the aversive experience is the conditioned stimulus (CS, e.g. movements and activities), and the fear it subsequently elicits is the conditioned response (CR). People might learn CS-US associations by actual experience, but also through receiving verbal threat information and by observing someone else experiencing a CS-US association (modelling). In general, fear conditioning is an adaptive form of learning, but it can become pathological when fearful responses to a CS persist, while the CS-US contingency is no longer present. For example, in the case of pain-related fear, an individual may still experience pain, but when no injury or damage (US) is present anymore, it does not help to avoid movements or activities (CS). Various tools to assess pain-related fear are available. Questionnaires can help to specify the patient's fear; does the patient expect damage, (re) injury, increasing pain, loss of a job or loss of independence (USs)? Pictorial stimuli of movements and activities help to assess which different stimuli (CSs) evoke fear and which specific expectations the patient has regarding a specific activity. The PHODA (Photograph Series of Daily Life Activities) is an instrument that is used for this purpose within exposure in vivo. An interview is used to gather information about the interrupting, interfering and identity-distorting effects of chronic pain, and the associations amongst them. After the assessment, when it is concluded that pain-related fear plays an important role in maintaining disability, exposure in vivo treatment is preferred. During exposure in vivo treatment behavioral experiments are prepared in which the patient is exposed to the CS (in case of pain-related fear to movements or activities), so that he can experience that his expectation regarding the occurrence of the US (the feared consequence such as damage or excruciating pain) is incorrect. This process is referred to as extinction. However, extinction does not lead to the forgetting or 'unlearning' of the CS-US association; a new, 'safe' association comes to exist alongside the old, "threatening" association. The CS thus becomes an ambiguous stimulus. Subsequently, in this chapter, scientific evidence obtained from other anxiety disorders is summarized and applied to the exposure treatment for pain-related fear. All components of treatment focus on learning that movements and activities are safe. This provides a patient with a new prediction that activities do not cause damage, although pain can occur. For example, education is given about the fear-avoidance model so that the patient can understand how his pain and disability are maintained. After this, the behavioral experiments will provide the patient with actual experiences violating the idea that painful movements are harmful. This chapter also discusses the effectiveness of exposure in pain-related fear, and identifies topics that have not yet sufficiently been studied scientifically, such as the role of pain-related fear in the transition from acute to chronic pain; other treatments than exposure in vivo that may also reduce pain-related fear, such as Acceptance and Commitment Therapy (ACT); the role of modelling within exposure and the application of exposure by non-psychologists; and generalization of exposure. Also, the role of safety behavior and whether exposure should focus on fear reduction or facilitating inhibitory learning has not yet been sufficiently investigated for pain-related anxiety. Addition of interoceptive exposure and paying attention to the personal goals of the patient could further enhance the effects of exposure for pain-related fear. The results of an experimental study on the role of information on the threat value of pain in healthy individuals are presented in chapter 3. Studies with the Voluntary Joystick Movement (VJM) paradigm confirmed that healthy participants can acquire fear for a painful movement in an experimental setting. The VJM paradigm uses a differential proprioceptive fear conditioning paradigm: one movement (Conditioned Stimulus, CS+) is repeatedly paired with a painful stimulus (Unconditioned Stimulus, Pain-US), and another (CS-) movement is not. In response to the CS+, defensive fear/avoidance responses (CRs) emerged. Defensive responses may arise from direct experience with a CS-US relation, but individuals also learn from verbal information and observation. The current study examined whether the acquisition of pain-related fear through direct experience can be intensified or weakened by verbally and visually transmitted information about the meaning of pain. A (bogus) sensor was attached to all participants' skin at the start of the experiment. To determine the intensity of the pain stimulus, participants received pain stimuli of increasing intensity. In this phase, the administration of these pain stimuli is not yet associated with movements. After the calibration phase, the researcher inspected the skin and told participants that an increased skin reactivity was observed. Therefore, 'skin reactivity measurement' would be shown during the task, which reflects the fragility of the skin for damage, such as blisters or burns. Participants were asked to carefully monitor the information from the sensor on the computer screen during the execution of the task. Participants saw a bar on the computer screen, varying in color from green via orange to red, containing a pointer that indicated the fragility of the skin. At the start of the task, the pointer was orange for all participants. Depending on the randomly assigned condition, the visual information differed during the task: participants in the inflation condition saw the pointer shift from orange to almost red, informing the participant about an increasing threat of skin damage. In the deflation condition, the pointer gradually shifted to almost green, informing the participants that the skin was safe. In the stable condition, the pointer remained in the orange zone throughout the experiment, keeping the threat to the skin equal with the previously given information. During extinction, the pointer remained in the red zone for participants in the inflation condition, in green for the deflation condition and in orange for the stable condition. In addition, the return of anxiety after a reinstatement procedure was investigated: after the extinction phase, two painful stimuli were given to half of the participants in each condition (reinstatement group), and not to the other half (control group). Participants' verbal reports and physiological measures showed acquisition and extinction of experimentally induced fear of movement-related pain. However, no effect of the threatening information about skin reactivity (inflation, deflation or stable) was found on acquisition, extinction or return of fear. However, a sensitization effect was observed: although the pain stimuli provided were of identical intensity, participants experienced them as increasingly painful over the course of the experiment. This suggests that the information about the fragility of skin may have increased the perceived threat in all groups. Furthermore, it was found that the fear of the painful movement returned in the reinstatement group after the administration of two additional pain stimuli post-extinction (a differential return of anxiety). In the physiological measurements (eye-blinking reflex), this group showed a (non-differential) return of fear for both the painful and the non-painful movement. It did not make a difference whether the participant was faced with an increasing, decreasing or stable threat during the task. The results suggest that the threat manipulation might not have worked or that it was not sensitive enough to yield group-specific effects. We replicated acquisition, extinction, and return of experimentally conditioned fear of movement-related pain, but the visual and verbal threat manipulation failed to generate additional effects. In chapter 4, a study on the generalization of extinction after exposure in vivo treatment is presented. Exposure in vivo has been shown to be a successful treatment in reducing pain-related fear, disability and experienced pain in chronic pain patients with pain-related fear. Experimental studies, however, show that extinction learning is context-dependent; there is a difference between learning in the treatment context, and apply what was learned in the home or work situation. This raises doubts whether post-treatment patients can generalize the extinction of pain-related fear for movements or activities that were performed during treatment to new activities that they may re-interpret as threatening. This study examined whether generalization to new threatening situations is promoted by exposure treatment (15 sessions) exposing patients to a variety of activities (Multiple Exposure condition), compared to a treatment exposing patients to the same set of activities several times in a row (Repeated Exposure condition). In the Multiple Exposure condition (N=4), patients were exposed once to at least 15 activities. In the Repeated Exposure condition (N=4), patients were exposed to only three activities during five sessions each. Generalization of extinction to new activities was tested by an independent observer who asked patients to perform several new activities, directly after treatment and at 6 months follow-up. Patients scored these activities (that were not addressed during treatment) as very threatening before the start of treatment. In addition, randomized replicated single case experimental designs were used to evaluate the effects of treatment on pain-related fear, personal relevant activities and experienced pain. Included were patients with CRPS-I reporting high levels of pain-related fear. It was hypothesized that the Multiple Exposure condition would facilitate generalization, but patients from both conditions performed equally well at both generalization tests. All patients performed the requested activities both immediately after treatment and 6 months after completion of treatment. Daily measures showed that the Multiple Exposure condition is preferred to reduce pain-related fear, pain catastrophizing and experienced pain. Patients in both conditions indicate that they can perform relevant personally activities again after treatment. This was also determined by an independent observer. Almost all patients showed a reliable improvement on the Leeds Reliable Change Index with respect to perceived disability, fear of (re-)injury, catastrophizing about pain and the threat value of activities, both after treatment and at 6 months follow-up. In conclusion, patients dare to perform new activities that were not performed during treatment after exposure in vivo treatment, irrespective of whether they have performed only a few activities during treatment, or many different activities only once. To improve other outcome measures, such as pain-related fear, catastrophizing about pain, and experienced pain, exposure to different activities is to be preferred over exposure to repeated activities. Chapter 5 presents a randomized controlled trial comparing exposure in vivo with pain-contingent physical therapy ("care as usual"), for CRPS-I patients with at least moderate levels of self-reported pain-related fear. CRPS-I highly affects patients' ability to perform daily life activities and pain-related fear might be a key target in order to reduce their disability levels and improve their quality of life. Treatments such as pain-contingent physical therapy, aiming at reducing pain have shown little improvements on pain and disability in CRPS-I, whereas novel exposure-based treatments targeting pain-related fears have shown to be promising. Therefore, a randomized controlled trial was conducted including 46 patients. The primary outcome was self-reported disability, for upper and lower extremity respectively. Secondary outcomes were self-reported pain-intensity, pain-catastrophizing, perceived harmfulness of physical activity, and health-related quality of life. Patients in both treatment conditions were offered 17 hours of treatment over a 17-week period. The results were striking. Exposure in vivo was superior to pain-contingent treatment as usual in reducing upper extremity disability from pre- to post-treatment and from pre-treatment to 6-months follow-up. Exposure in vivo was also superior in reducing lower extremity disability from pre-treatment to 6-months follow-up, but not from pre- to post-treatment. Also, more patients from the exposure condition showed reliable progression on the Leeds Reliable Change Index than patients who received pain-contingent physiotherapy. All secondary outcomes significantly favored exposure pre- to post-treatment, as well as pre-treatment to 6-months follow-up: self-reported pain-intensity, pain-catastrophizing, perceived harmfulness of physical activity, and health-related quality of life. Concluding, exposure to daily activities shows to be more effective than a protective pain-contingent treatment in reducing self-reported disability in daily life, as well as in reducing pain-related fear, experienced pain and in improving quality of life, for CRPS-I patients with at least moderate levels of pain-related fear. A remaining question is if exposure is cost-effective as well; for example in reducing the use of health-care or increased participation in work. Therefore, chapter 6 evaluates the cost-effectiveness of exposure in vivo in patients with Complex Regional Pain Syndrome Type I (CRPS-I) and elevated levels of pain-related fear, compared to pain-contingent physical therapy. Data from the randomized controlled trial presented in chapter 5 were used to compare the cost-effectiveness of exposure in vivo versus pain-contingent physical therapy from a societal perspective. Interventions such as exposure in vivo and physiotherapy influence the quality of life of patients. Improvements in quality of life can justify the costs of an intervention. An effect on the quality of life can be regarded as the 'usefulness' or utility of the intervention, and the value of this must outweigh the costs. In order to operationalize utility, the outcome measure 'Quality-Adjusted Life Years' (QALY) is used. The main outcomes were changes in the SF-36 physical component scale and quality-adjusted life years. To calculate costs, intervention costs, other healthcare costs, other costs to patient and family and productivity losses were included. Patients completed cost diaries every month, up to six months after the end of treatment. In economic evaluations, it is common to determine the degree of uncertainty in costs and effects. Uncertainty was estimated using non-parametric bootstrap analysis, cost-effectiveness acceptability curves and cost-effectiveness planes. Over 6 months after treatment, exposure in vivo resulted in greater improvement in physical health-related quality of life and quality adjusted life years than pain-contingent physical therapy. Despite higher initial treatment costs, healthcare costs were significantly reduced, and exposure in vivo showed a tendency to reduce total societal costs compared to pain-contingent physical therapy. Patients used significantly less medication after exposure in vivo than after pain-contingent physical therapy (medication on prescription as well as over-the-counter medication). They also need significantly less help from family members. The results of the bootstrap analyses showed that in 95%, exposure gives significantly better effects at lower costs for the physical health-related quality of life and the quality-adjusted life years. Sensitivity analyses, one with different treatment costs and one with complete cases only, confirmed robustness of these findings. Concluding, exposure in vivo is more cost-effective than PPT in CRPS patients with pain-related fear. The initial higher costs for exposure in vivo, explained by the additional costs for a psychologist, are offset by a significant long-term reduction of costs for health care utilization. Furthermore, tendencies to reduce hours absent from work and total societal costs were found. The results justify reimbursement of exposure treatment by insurance companies and incorporation of exposure in vivo as a viable treatment for patients with CRPS-I and pain-related fear in national and international clinical guidelines. In chapter 7 provides a general discussion, summarizing the main findings of the studies. In addition, recent developments are discussed that are related to the topics described in this thesis. Since the conduction of the studies in this thesis, insights in the mechanism of exposure treatment have evolved. The emotional processing theory presumed that habituation was the mechanism by which exposure was effective: repeated confrontation with a feared stimulus was supposed to result in a decrease in fear. Recently, it became clear that fear-reduction is not a reliable indicator of learning during exposure. This led to a changed perspective on the underlying mechanism of exposure, which is currently regarded as a method to disconfirm negative, fearful expectancies by experiencing disconfirmation between the expectancy and experience (expectancy violation). Due to the learning of new CS-US associations, the meaning of the CS becomes ambiguous; besides the old excitatory meaning the CS (predicting the US) now also holds an inhibitory meaning as well (predicting absence of the US). However, in patients with elevated anxiety levels (due to anxious personality traits or anxiety disorders), this inhibitory learning mechanism may be hampered. Techniques to optimize inhibitory learning have been described, and several techniques can be easily integrated in exposure for patients with pain-related fear. For example, designing more "extreme" behavioral experiments that are likely to provide a pronounced mismatch between expectancy and experience, instead of stepwise following a fear hierarchy. In addition, patients should be instructed to focus on the non-occurrence of the feared outcome (no mental distraction, no talking during the performance of threatening activities). Exploring alternative outcomes before actual performance of activities might reduce the surprise-value of the outcome of a behavioral experiment, so formulating alternative hypothesis and showing the activity by the therapist before performance of the patient is currently omitted. Exposure can easily be provided in multiple contexts, with varying levels of fear, and patients might be advised to mentally reinstate the exposure-context. Affect labeling, verbalization of one's current emotional experience, is another technique to optimize inhibitory learning. Activities in an exposure treatment for pain-related fear are and should be primarily chosen based on the mismatch they create between expectancy and actual experience, but personal relevance and likeability might also be important, although this subject has not been extensively studied yet. Recent adaptations to the fear-avoidance model suggest that when patients prefer control over pain, they tend to avoid activities. When patients have other goals, such as accomplishing personal relevant tasks, this results in confrontation with painful activities. Cross-sectional studies show that these goal-conflicts experienced by chronic pain patients (between avoidance behavior aimed to prevent harm or pain, and confronting activities, helping them to engage in valued daily activities) are associated with pain-related fear, higher pain-intensity and negative affect. Furthermore, compliance with exposure treatment might be enhanced when a patient experiences immediate reward in the first sessions. Finally, directions for future research are discussed. Besides integrating the insights on inhibitory learning that are previously addressed, it might be worthwhile to implement additional techniques to exposure in vivo. For example, counterconditioning might help to restore the affective value of activities that are not avoided any more after exposure but might still be disliked because of their previous negative association with pain and/or harm. Another direction to improve the effects of exposure is the addition of pharmacological enhancers; agents that might enhance inhibitory learning and memory-processes during exposure treatment. Furthermore, it seems also important to strengthen positive, protecting mechanisms in chronic pain patient, for example resilience and acceptance. Based on the abovementioned findings, it is concluded that exposure in vivo is successful in reducing disability, pain-related fear and the pain-experience in patients with CRPS-I. Furthermore, this novel behavioral treatment reduces healthcare costs. It seems to be preferred to include multiple activities in an exposure in vivo treatment for pain-related fear, creating multiple mismatches between the negative outcome a patient is expecting, and the actual experience of performing the activity without the expected negative consequences. Techniques to optimize inhibitory learning are recently implemented, and future studies specifically in pain-related fear should examine if these further improve the effects of treatment for fearful chronic pain patients. status: published