| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
REVIEW ARTICLES |
From the PRIDE Research Foundation (J.D., P.B.P.), Dallas, Texas; and Departments of Psychiatry and Rehabilitation Sciences (R.J.G.), The University of Texas Southwestern Medical Center, Dallas, Texas.
Address reprint requests to: Robert J. Gatchel, PhD, Division of Psychology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, Texas 75390-9044. Email: robert.gatchel{at}utsouthwestern.edu
ABSTRACT
OBJECTIVE: Numerous studies have documented a strong association between chronic pain and psychopathology. Previous research has shown that chronic pain is most often associated with depressive disorders, anxiety disorders, somatoform disorders, substance use disorders, and personality disorders. The primary objective of this review article is to describe the nature of the relationship between chronic pain and each of these types of psychopathology. In addition, this article will explore how each of these disorders are expressed within the context of chronic pain, with a consideration of both diagnostic and treatment issues.
METHODS: Medline and PsychLit searches of the chronic pain/psychopathology literature from 1980 through 2000 were conducted using the keywords chronic pain, psychopathology, psychiatric disorders, and psychological disorders.
RESULTS: The relationship between chronic pain and psychopathology has generated substantial empirical and theoretical interest, with depressive disorders receiving much of the attention.
CONCLUSIONS: Although no single theoretical model can fully explain the causal relationship between chronic pain and psychopathology, a diathesis-stress model is emerging as the dominant overarching theoretical perspective. In this model, diatheses are conceptualized as preexisting, semidormant characteristics of the individual before the onset of chronic pain that are then activated and exacerbated by the stress of this chronic condition, eventually resulting in diagnosable psychopathology.
Key Words: chronic pain, • psychopathology, • psychiatric disorders, • psychological disorders.
Abbreviations: CLBP = chronic low back pain;; DSM = Diagnostic and Statistical Manual of Mental Disorders;; MDD = major depressive disorder;; MMPI = Minnesota Multiphasic Personality Inventory;; PD = Personality Disorder;; SNAP = Schedule for Nonadaptive and Adaptive Personality;; SCID = Structured Clinical Interview for the DSM.
Chronic pain is an increasingly costly and debilitating medical condition in industrialized countries. As more has been learned about this condition, the importance of psychological and social factors in understanding chronic pain has become increasingly recognized. Chronic pain is now widely viewed as a biopsychosocial phenomenon, in which biological, psychological, and social factors dynamically interact with one another (1). As psychological (ie, behavioral, cognitive, and affective) factors have been explored, it has become increasingly evident that chronic pain is associated with high rates of diagnosable psychopathology.
There are several reasons why it is important to identify psychopathology in chronic pain patients. Most importantly, unrecognized and untreated psychopathology can significantly interfere with successful rehabilitation of these patients (2). Rehabilitation programs without an adequate mental health component may therefore be "doomed to failure" (2). Psychopathology may also increase pain intensity and disability, thus serving to perpetuate pain-related dysfunction (3). For example, anxiety has been found to decrease pain threshold and tolerance (4); anxiety and depression have been associated with magnification of medical symptoms (5); depression has been associated with less successful treatment outcomes (6); and emotional distress has been linked to physical symptoms through autonomic arousal, vigilance, and misinterpretation (7), or somatic amplification (8).
Before proceeding, it should be noted that the term chronic pain includes a large variety of heterogeneous clinical conditions (9). The relationship between psychopathology and chronic pain may be different for distinct clinical conditions; it may also differ depending on the clinical context. For example, psychosocial reactions to chronic pain associated with advanced malignancy may not be similar to reactions to benign chronic pain in which there is an expectation of improvement. In this review, we focus primarily on chronic, nonmalignant musculoskeletal pain. When appropriate, we extend the discussion to other types of chronic pain conditions.
HISTORICAL OVERVIEW
Research on psychopathology in the chronic pain population, conducted mostly on patients with chronic low back pain (CLBP) in the 1980s, documented increased prevalence of depression, anxiety, substance abuse/dependence, "somatization," and personality disorders compared with the general population (10–13). For example, rates of major depressive disorder (MDD) ranged from 34% to 57% in these studies, compared with rates of 5% to 26% in the general population (14). These studies were an important first step in documenting high rates of psychopathology in chronic pain patients. However, a number of methodological flaws have limited their usefulness. For example, several of the studies failed to use the DSM (Diagnostic and Statistical Manual of Mental Disorders), the standard nomenclature of psychiatric illness. Other studies used diagnostic criteria consistent with this nomenclature, but diagnostic decisions were made on the basis of self-report questionnaires or loosely structured psychiatric interviews (15).
Methodological Refinements
More recent studies have addressed the limitations of earlier ones by using semistructured clinical interviews based on DSM diagnostic criteria. As early as 1983, Reich et al. (13) recommended the DSM as the most accurate means for classifying patients with chronic pain in a uniform manner. The multiaxial classification format allows users to consider both physiological and psychological components of chronic pain in a systematized manner.
Structured clinical interviews based on DSM criteria, when compared with self-report questionnaires or loosely structured interviews, represent a further methodological refinement by facilitating direct comparisons of rates of psychopathology across different studies (16). One such interview that has come to be used in studying psychopathology in chronic pain patients is the Structured Clinical Interview for DSM-III-R (17). An important advantage of the SCID, in addition to its good reliability and promising validity, is that it allows the determination of current and lifetime diagnoses. It also makes it possible to determine whether the current pain episode was present before or after the occurrence of psychopathology (2). By clarifying the temporal relationship between the onset of pain and the onset of diagnosable psychopathology, such data should aid in answering the "chicken-and-egg" question (18) of which comes first, the pain or the psychopathology.
Clinical Research Issues
In a study evaluating one aspect of this chicken-and-egg question, CLBP patients (N = 200) were assessed for current and lifetime psychiatric disorders using the SCID to make DSM-III-R axis I and axis II diagnoses (19). Results showed that even when the somewhat controversial category of somatoform pain disorder was excluded, 77% of patients met lifetime diagnostic criteria and 59% demonstrated current symptoms for at least one psychiatric diagnosis. In comparison, lifetime and current rates of mental disorders in the general population were estimated to be 29% to 38% and 15%, respectively (14, 20, 21). The most common diagnoses in the Polatin et al. study (19) were MDD, substance abuse, and anxiety disorders. In addition, 51% met criteria for at least one personality disorder. Several gender differences were observed, with females more likely to have a lifetime or current diagnosis of MDD and a lifetime anxiety disorder diagnosis and males more likely to have a current substance abuse diagnosis. The overall prevalences were significantly greater than the base rates for the general population, supporting the earlier research discussed above. Furthermore, of those patients with a lifetime history of a psychiatric disorder, 54% of those with MDD, 94% of those with substance abuse, and 95% of those with anxiety disorders had experienced these symptoms before the onset of their back pain. These results were the first to demonstrate that certain psychiatric disorders appear to precede chronic pain (substance abuse, anxiety disorders), whereas others (specifically MDD) develop either before or after the onset of pain. Although these findings increase our understanding of the relationship between psychiatric disorders and chronic pain, at least for CLBP patients, the retrospective design of this study does not allow us to answer the question of whether psychiatric disorders in chronic pain patients are consequences of experiencing the chronic pain or whether preexisting disorders act as predispositions for pain to become chronic (22).
A more recent study addressed the chicken-and-egg question in a different manner. CLBP patients (N = 421) were evaluated within 6 weeks of acute back pain onset with a standard psychosocial assessment battery that included the SCID (23). The purpose of this study was to evaluate the predictive power of a psychosocial assessment battery in determining which patients presenting for acute back pain will go on to develop chronic pain disability (as measured by return-to-work status at a 1-year follow-up evaluation). Among the findings of this study were that major psychopathology was not predictive of the development of chronic pain disability. Although the results suggested that psychopathology does not directly precede the onset of chronic pain disability, a definitive answer to the chicken-and-egg question remained elusive.
Another relevant issue that has been investigated is the prevalence of psychiatric disorders in acute vs. chronic pain. Kinney et al. (24) examined this issue in acute vs. CLBP. They found higher rates of psychopathology in the CLBP group. In particular, the CLBP patients had higher rates of MDD, substance use disorders, and personality disorders than the acute LBP patients. In contrast, the acute pain patients were diagnosed with more anxiety disorders. This pattern of results, which was also found in a study of acute and chronic pain patients with temporomandibular disorder (16), indicates that higher rates of psychiatric disorders are not solely related to the onset of pain per se but are linked to the development of chronicity. Furthermore, these results lend support to Gatchel’ s model of the progression from acute pain to chronic pain disability, in which anxiety is considered to be a common reaction to acute pain with more disabling and varied psychopathology associated with chronic pain (2, 18).
Gatchel (18) developed a three-stage conceptual model to further our understanding of the transition from acute pain to chronic pain disability and accompanying psychosocial distress. As pain becomes increasingly chronic, Gatchel indicates that patients undergo significant psychological changes, producing "a layer of behavioral/psychological problems over the original nociception or pain experience itself" (2). Stage 1 of Gatchel’ s model is associated with normal emotional reactions such as fear, anxiety, and worry as a result of the perception of pain during the acute phase. If the pain persists beyond what is normal for acute pain (2–4 months), the individual enters stage 2. This stage is associated with a wider array of behavioral and psychological reactions and problems, such as learned helplessness, distress, anger, and somatization, that are the consequence of suffering with the more chronic nature of the pain. Gatchel hypothesized that the form these problems take depends largely on the premorbid or preexisting personality/psychological characteristics of the individual as well as current socioeconomic factors and other environmental conditions. This stage of the model reflects a diathesis-stress perspective, in which the stress of coping with chronic pain exacerbates the individual’s preexisting characteristics (diatheses).
Gatchel (18) further hypothesized that as these reactions and problems persist, the individual’s life begins to completely revolve around the pain as a result of the chronic nature of the problem. The individual then enters stage 3, which can be viewed as the acceptance or habituation to some aspects of a "sick role." The sick role excuses patients from their normal responsibilities and social obligations, which may become a potent reinforcer for not becoming healthy.
Although Gatchel’s model has added conceptual clarity to the relationship between chronic pain and psychopathology, additional research has focused on determining whether psychiatric disorders are a limiting factor in the successful rehabilitation of patients with chronic pain. Gatchel et al. (25) used the SCID to assess the prevalence of current and lifetime DSM diagnoses in a sample of 152 CLBP patients beginning an intensive 3-week rehabilitation program. These patients were then followed over time, with treatment outcome being defined as return-to-work status 1 year after program completion. Despite high rates of axis I and axis II psychiatric disorders in this sample, neither type nor degree of psychopathology were found to be predictive of a patient’s ability to successfully return to work.
Recent studies of chronic pain and psychopathology have extended the existing body of research by investigating whether changes in psychopathology occur after intensive rehabilitation of CLBP patients (26, 27). Owen-Salters et al. (26) used the SCID to evaluate 56 patients for current psychiatric disorders on admission to a comprehensive 3-week rehabilitation program and again at 6 months after completion of the treatment program. The results documented significant decreases in prevalence of psychiatric disorders, particularly somatoform pain disorder and MDD. In a methodologically similar study, decreased prevalences were found of axis II personality disorders 6 months after completion of the treatment program in a sample of 105 CLBP patients (27). These two studies demonstrate that effective rehabilitation can significantly decrease the high rates of psychiatric disorders found in CLBP patients, although alternative explanations for the decreased rates of axis II disorders in the Vittengl et al. (27) study have also been proposed and will be discussed below.
It should be noted that anger is a common, and perhaps the most prominent, affective response to chronic pain (28). However, the current version of the DSM (29) largely fails to address anger as a mental disorder. As future versions of the DSM begin to address this issue, anger disorders are expected to be a frequent diagnosis in this population.
Unresolved Issues
The research findings reviewed above have answered many questions regarding psychiatric disorders and chronic musculoskeletal pain. However, important unresolved issues remain, including the precise nature of the causal relationship between chronic pain and psychopathology. It should be noted that research investigating the relationship between chronic pain and psychopathology has not been limited to chronic musculoskeletal pain patients. High rates of psychiatric disorders have also been found to be associated with headaches (30), temporomandibular disorders (31), pelvic pain (32), and fibromyalgia (33). Furthermore, it has become increasingly evident that medical conditions in general, and particularly chronic ones, are associated with higher rates of psychiatric disorders (34). For example, Katon and Sullivan (35) found that between 15% and 33% of medically ill inpatients were suffering from mood or anxiety disorders, compared with 2% to 4% of the general population. High rates of medical and psychiatric comorbidity have prompted some theorists to develop models of the relationship between these types of disorders (36). These models seem promising but have yet to be validated empirically. In the chronic pain literature, with the exception of Gatchel’s model (2, 18), general theoretical models addressing these issues are lacking. However, recent theoretical advances have been made regarding the nature of the relationship between chronic pain and depression (37). These advances will be covered in more detail in the next section of the literature review, which deals specifically with chronic pain and depression.
CHRONIC PAIN AND DEPRESSION
Beginning with the influential article of Romano and Turner (38), the association between chronic pain and depression has generated more research and theoretical interest than any other area of the chronic pain/psychopathology literature. Much of this interest can be attributed to the frequency with which chronic pain patients suffer from depression (39). A number of the studies (19, 24) reviewed above identified extremely high rates of MDD in chronic pain patients, with current and lifetime rates of this disorder of about 45% and 65%, respectively, in the CLBP population and both current and lifetime rates of about 80% in the chronic upper extremity pain population. However, most studies report prevalences in the moderately high range. For example, Banks and Kerns (37) reviewed 14 studies that used DSM criteria for diagnosing depression in chronic pain patients and found that 9 of these studies reported current prevalence of MDD to be between 30% and 54%. In contrast, recent estimates of current and lifetime major depression for the entire US population are 5% and 17%, respectively (40).
Diagnostic Issues
Despite wide agreement that depression is frequently associated with chronic pain conditions, several problematic issues have proven to be barriers to research in this area (39). The first is the definition of depression itself. The term "depression" has been used to refer to a mood, a symptom, and a syndrome (37, 39). As a result, depression has been assessed in a variety of ways, including self-report instruments, projective tests, chart reviews, and structured and unstructured clinical interviews. The definitional and assessment variability have made it difficult to compare the results of different studies. Another type of variability that has proven to be a barrier to research in this area has to do with the types of patients being evaluated (39). Many studies have examined heterogeneous samples of chronic pain patients, that is, patients with a variety of pain sites and/or patients seen in different treatment settings. However, some researchers have suggested that certain pain sites are more commonly associated with depression than others (41) and that the incidence of depression may vary across treatment settings (42). One additional problem in studying the relationship between depression and chronic pain is "criterion contamination" (43). This term refers to the overlapping symptomatology of chronic pain and depression. Because the diagnostic criteria for MDD include several somatic symptoms that can also be attributed to chronic pain (eg, sleep disturbance, motor retardation, loss of energy, and change in appetite and weight), diagnosing depression in this population is not always straightforward. This problem may not be limited to diagnosing depression in chronic pain patients but may extend to the medically ill population in general (44). Criterion contamination may be particularly problematic with psychological instruments assessing depression, like the Beck Depression Inventory, which were standardized on psychiatric populations from which those with significant physical illness and disability were excluded (45).
Banks and Kerns (37) suggest that, until future research determines appropriate diagnostic criteria and measures of depression among chronic pain patients (see Ref. 46 for a proposal about such criteria) and the medically ill, standardized diagnostic systems such as the DSM are the most reliable means of producing valid diagnoses of MDD. Use of the DSM will also address the aforementioned problems of definitional and assessment variability (37). As discussed above, use of a structured clinical interview based on DSM criteria, such as the SCID, will further reduce assessment variability and facilitate direct cross-study comparisons.
Is There a Causal Relationship?
Despite significant barriers to answering research questions, recent progress has been made, particularly in terms of understanding the causal nature of the relationship between chronic pain and depression. In terms of the temporal relationship between these disorders, depression has been shown to be an antecedent to chronic pain (19), a consequence of chronic pain (19, 47), and a concomitant biological relative of chronic pain (48).
Fishbain et al. (49) reviewed these and other studies and found that five major hypotheses have been proposed concerning the timing and causal relationship of depression to chronic pain. These have been termed the 1) antecedent hypothesis—depression precedes the development of chronic pain; 2) consequence hypothesis—depression is a consequence and follows the development of pain; 3) scar hypothesis—episodes of depression occurring before the onset of pain predispose an individual to a depressive episode after pain onset; 4) cognitive behavioral mediation hypothesis—cognitions mediate the relationship between chronic pain and the development of depression; and 5) common pathogenetic mechanisms hypothesis.
Fishbain et al. (49) reviewed 40 studies either directly or indirectly addressing these hypotheses. These authors found that the large majority of studies addressing the antecedent hypothesis failed to support it. In contrast, they found that the consequence hypothesis was supported by each of the relevant 15 studies they reviewed. The cognitive behavioral mediation hypothesis was supported by five of six studies reviewed. They indicate that this latter hypothesis is not necessarily incompatible with the consequence hypothesis and note, in fact, that the direction of the relationship of the disorders was the same for all five studies, from pain to depression, thus adding further support to the consequence hypothesis. The scar hypothesis, which assumes a genetic predisposition to recurrent major depression, received partial support, with some studies showing a higher percentage of chronic pain patients who have first-degree relatives with depressive spectrum disorders than is found in the general population.
Fishbain et al. (49) did not locate any studies directly addressing the common pathogenetic mechanisms hypothesis. However, others have suggested that pain and depression are physiologically identical, considering chronic pain to be a variant of depression (50, 51) or, in a parallel construction, a "masked" depression (52). The physiological similarities between chronic pain and depression have been reviewed (39). For example, both nociceptive and affective pathways coincide anatomically. Furthermore, norepinephrine and serotonin, the two neurotransmitters most implicated in the pathophysiology of mood disorders, are also involved in the gate-control mechanism of pain. Finally, antidepressants have been found to be effective in the treatment of some patients with chronic pain. Despite evidence for common pathogenetic mechanisms, there are a number of reasons to regard them as distinct disorders, including nosological clarity, the fact that each condition exists without the other for many patients (39), and the finding that the onset of pain and depression do not typically coincide.
The Diathesis-Stress Model
Returning to the Fishbain et al. (49) study, these authors concluded that the consequence hypothesis is most strongly supported, with additional support for the cognitive-behavioral mediation and scar hypotheses. A recent theoretical proposal by Banks and Kerns (37), when placed within the framework of Fishbain et al.’ s study, may help to further clarify the relationship between chronic pain and depression. After demonstrating that rates of depression seem to be higher in populations of patients with chronic pain than in populations of patients with other chronic medical conditions, they proposed a diathesis-stress model of the development of depression in chronic pain patients that is consistent with the consequence, scar, and cognitive behavioral mediation hypotheses. In their model, the diathesis is psychological in nature and could include negative schemas (53); internal, stable, and global attributions leading to learned helplessness (54); and/or deficits in instrumental skills (55). These diatheses are conceptualized as preexisting, semidormant characteristics of the individual before the onset of chronic pain that are then activated by the stress of this chronic condition. The stress component of the model proposed by Banks and Kerns (37) refers to the nature of the chronic pain experience. They suggest that chronic pain is more likely to result in depression than other chronic medical conditions because of the uniquely challenging nature of stressors associated with chronic pain, including the aversive sensory and distressful emotional aspects of the pain symptoms, impairment and disability, secondary losses that occur across various domains (eg, work and social relationships), and perceived invalidating responses from the medical system. The last stressor has recently been reviewed in some detail by May et al. (56), who point out that the disparity between expressed symptoms, pathological signs, and perceived disability in CLBP has led to the moral character of the sufferer forming a constant subtext to medical discourse about the condition. It should be noted that the Banks and Kerns (37) model is an elaboration of Gatchel’s more general model (18) applied specifically to the relationship between depression and chronic pain.
The model proposed by Banks and Kerns (37) is obviously consistent with the consequence hypothesis. It is also consistent with the cognitive-behavioral mediation hypothesis in that cognitions (negative schemas, attributions) and behaviors (instrumental skills) mediate the relationship between pain and depression. Finally, if the scar hypothesis could be broadened to include not only a genetic predisposition to recurrent major depression but also to include psychological predispositions such as negative schemas or a depressive attributional style, then the diathesis-stress model would also be consistent with this hypothesis. Empirical support for the diathesis-stress model was recently obtained (57) with the finding that each of the three cognitive/behavioral variables outlined by Banks and Kerns emerged as unique predictors of self-reported depression in CLBP patients. Moreover, consistent with the cognitive-behavioral mediation hypothesis, when these three variables were held constant, pain and disability did not have a significant association with self-reported depression.
The similarities between the models proposed by Banks and Kerns (37) and, earlier, Gatchel (2, 18) are striking. Gatchel stated, " Many patients whose pain later becomes chronic may ‘bring with them’ certain premorbid or predisposing psychological/personality characteristics or disorders (a diathesis) that are exacerbated by the stress of attempting to cope with pain" (2). With a diathesis-stress theoretical framework emerging, future research will focus on identifying additional psychological diatheses, further understanding the stressful nature of the chronic pain experience, and, most importantly, clarifying the factors that mediate the relationship between diathesis/stress and the development of MDD and other forms of diagnosable psychopathology. Weisberg and Keefe (58) have also noted the important heuristic value of such a diathesis-stress model in better understanding chronic pain.
CHRONIC PAIN AND SUBSTANCE USE DISORDERS
Numerous studies have identified high prevalence rates of substance use disorders in patients with chronic pain (11, 13, 19, 59). These and other relevant studies were reviewed by Brown et al. (60). Of the studies using standard measures for diagnostic assessment, including the SCID, these reviewers found that rates of current substance use disorders ranged from 15% to 28%, whereas lifetime rates ranged from 23% to 41%. Lifetime prevalences are clearly higher than the 16.7% found in the general population (61), whereas current prevalence rates are "probably" higher (62). Substance use disorders are more prevalent for males than females, both in the general population (61) and in the chronic pain population (60).
Turning to the issue of the relationship between chronic pain and substance use disorders, the research literature suggests that a variety of temporal and causal pathways are likely. Polatin et al. (19) found that 94% of the chronic pain patients with lifetime substance use disorders experienced the onset of these disorders before the onset of their chronic pain. They suggested that chronic pain may not precipitate substance use disorders as often as many clinicians believe. Others have speculated that premorbid substance use disorders may increase the risk for traumatic illness and work-related accidents, a certain proportion of which will develop into chronic pain conditions (60). As mentioned earlier, rates of current substance use disorders are "probably" higher in chronic pain patients than in the general population. However, Brown et al. (60) found that chronic pain patients were no more likely to have a current substance use disorder than other patients in a primary care setting, suggesting that chronic pain is not associated with a unique risk for substance abuse. Despite these findings, Brown et al. found that chronic pain patients are at an increased risk for new substance use disorders during the 5 years after the onset of chronic pain, as compared with other 5-year periods in their lives. The risk may be related to iatrogenic factors, with estimates of current addiction to narcotic analgesics ranging from 3% to 16% (63). This risk seems to be greatest in individuals with a previous history of substance abuse/dependence and in those with a childhood history of physical or sexual abuse, suggesting that these factors are premorbid (ie, prechronic pain) diatheses for new or recurrent substance dependence.
Additional important findings related to chronic pain and substance use disorders have also been discussed in the research literature. In terms of the relationship between substance use disorders and long-term treatment outcomes, Burton et al. (64) found that the presence of a preinjury substance use disorder in chronic upper extremity pain patients was associated with a lower return to work rate 1 year after rehabilitation, whereas Gatchel et al. (25) did not find such an association in CLBP patients. Fishbain et al. (65) examined the comorbidity of substance use disorders with other DSM disorders. They found that patients with a substance diagnosis had higher rates of MDD, anxiety disorders, and personality disorders than patients with no substance diagnosis, suggesting that these patients may be self-medicating psychiatric symptoms with drugs or alcohol.
Abuse vs. Dependence
Before proceeding further, it should be noted that substance use disorders are typically broken down into the categories of abuse and dependence. Substance abuse is defined in the DSM-IV as a maladaptive pattern of substance use leading to clinically significant impairment or distress, as manifested by at least one of several criteria (eg, recurrent substance use in situations in which it is physically hazardous). The diagnostic criteria for substance dependence involve both behavioral (or "psychological" ) dependence and physiological dependence. The criteria for behavioral dependence emphasize the substance-seeking activities and related evidence of pathological use patterns, whereas the criteria for physiological dependence emphasize the body’ s reaction to heavy and prolonged use of a substance (66).
The most commonly abused substances in chronic pain patients seem to be alcohol (current and lifetime) and narcotics (current). However, a controversy exists in the chronic pain literature about how to classify patients who are maintained on narcotic analgesic medications (63). Most of these patients will meet the criteria for tolerance and withdrawal, almost ensuring them a DSM-IV diagnosis of substance dependence. However, the overwhelming majority of these patients are not actually behaviorally or psychologically "addicted" to narcotic medications (63). In light of the unique issues relating to this patient population, the American Society for Addiction Medicine (ASAM) developed separate criteria for defining addiction in chronic pain patients treated with narcotics (63). These criteria are focused on symptoms of behavioral dependence, including the presence of adverse consequences associated with the use of narcotics, loss of control over the use of narcotics, and preoccupation with obtaining narcotics despite the presence of adequate analgesia.
The best method to identify substance use disorders in chronic pain patients has also been an ongoing issue. With substance-abusing individuals, in general, there is always the potential for denial, social desirability bias, and memory lapse (60). With chronic pain patients, there may also be fears that pain-relieving medications will be reduced or discontinued, that they will not be considered good candidates for pain rehabilitation treatment if the physician finds out about their patterns of medication and other substance use (63), and/or that this information will work against them in potential future worker’s compensation litigation (67). This latter study investigated the concordance of self-reported substance use and urine toxicology indicators of substance use in a sample of chronic pain patients. They found that 9% of the patients provided incorrect self-report information about current substance use, mostly about illicit substances (cocaine and cannabis). Chronic pain patients who incorrectly denied substance use were more likely to be younger, to have a worker’ s compensation case, and to have been assigned a DSM diagnosis of polysubstance abuse in remission. The authors suggest that this last variable may be attributable to chronic pain patients feeling comfortable in disclosing a past history of substance abuse, knowing that such disclosure would in no way affect their current situation. Fishbain et al. (67) indicate that young age, worker’s compensation status, and a diagnosis of polysubstance abuse in remission may indicate a need for urine toxicology in a chronic pain patient denying illicit drug use but in whom there is suspicion of such. Despite the concerns and factors outlined by Fishbain and others, there is widespread agreement that self-report is the most accurate method currently available for assessing substance use disorders (68).
CHRONIC PAIN AND ANXIETY DISORDERS
Several studies have documented high rates of anxiety disorders among chronic pain patients (19, 59, 64). Anxiety disorders include panic disorder, agoraphobia, specific phobia, social phobia, posttraumatic stress disorder, obsessive compulsive disorder, and generalized anxiety disorder. Although there is little consistency in the chronic pain research literature regarding specific anxiety disorder diagnoses, panic disorder and generalized anxiety disorder seem to be most commonly diagnosed (65). The nonanxiety disorder diagnosis of adjustment disorder with anxious mood has also been frequently used (59).
In studies that used a structured DSM clinical interview, such as the SCID, overall prevalence for anxiety disorders ranged from 16.5% to 28.8% (65). Although overall prevalences are close to those estimated for the general population, recently presented evidence suggests that anxiety disorders are more often associated with chronic pain than has been reported in the research literature (65). Studies that have made a distinction between lifetime and current anxiety disorders (19, 64) have found that lifetime prevalences are close to those estimated for the general population, whereas current prevalence rates are significantly higher in chronic pain patients. The relationship between anxiety disorders and long-term treatment outcomes has also been investigated. Burton et al. (64) found that the presence of an anxiety disorder (lifetime or current) in patients with chronic upper extremity pain was associated with a lower return to work rate 1 year after rehabilitation, whereas Gatchel et al. (25) did not find such an association in CLBP patients.
In several studies described earlier (16, 24), researchers found that although chronic pain patients had much higher rates of overall psychopathology than acute pain patients, anxiety disorders were diagnosed frequently in both groups. These data support Gatchel’s model (2, 18) of the progression from acute pain to chronic pain disability, in which anxiety is considered to be a common reaction to acute pain, with more disabling and varied psychopathology associated with chronic pain. The diathesis-stress model, described in the section above on depression, may also be applicable to the relationship between chronic pain and anxiety disorders. Polatin et al. (19) found that 95% of those diagnosed with anxiety disorders in their sample of work-disabled CLBP patients had experienced these disorders before the onset of pain, suggesting that a premorbid physiological or psychological diathesis exists that is then exacerbated by the stress of the chronic pain experience. One diathesis may be a genetic predisposition to panic disorder or one of the other anxiety disorders. Another diathesis may be anxiety sensitivity, which is considered to be one of three fundamental trait sensitivities that amplify an array of fear reactions and phobias according to the expectancy model of fear (69).
Psychophysiological Factors
Once an anxiety response is in place, chronic pain may be maintained or exacerbated through direct physiological mechanisms (1). Fear of pain and fear of movement/reinjury, both of which lead to further physical deconditioning through avoidance of activities that have potential for long-term pain reduction, may also contribute to the maintenance of pain (70). Avoidance, in turn, may be reinforced through operant learning mechanisms. For example, unwanted responsibilities may be avoided, or anxiety associated with the anticipation of further pain-related experiences may be reduced (71). Cognitive factors may also be involved, with avoidance leading to decreased self-efficacy and increased expectations that stimulation will increase pain, which in turn leads to increased avoidance. The result is a self-defeating cycle between cognition and behavior (70). Supporting evidence for these ideas has come from a study reporting that pain-related fear-avoidance beliefs about work are the most specific and powerful factors accounting for disability and work loss associated with CLBP (72). One additional cognitive factor that may be important in maintaining chronic pain is the tendency of anxiety-sensitive patients to catastrophically misinterpret sensations of arousal that are associated with pain (70).
CHRONIC PAIN AND SOMATOFORM DISORDERS
Perhaps the most common psychiatric disorder diagnosis that chronic pain patients receive is pain disorder, which is one of the somatoform disorders listed in the latest version of the DSM (29). According to DSM-IV, "The common feature of the Somatoform Disorders is the presence of physical symptoms that suggest a general medical condition and are not fully explained by a general medical condition, by the direct effects of a substance, or by another mental disorder" (29). Other somatoform disorders, including somatization disorder, conversion disorder, and hypochondriasis, are not frequently diagnosed in chronic pain populations.
Pain Disorder
Pain disorder is diagnosed when pain is the predominant focus of clinical presentation, when the pain causes significant distress or functional impairment, and when psychological factors are judged to have an important role in the onset, severity, exacerbation, or maintenance of the pain (29). The names and diagnostic criteria for pain disorder have evolved from the two most recent versions of the DSM. In DSM-III (73), psychogenic pain disorder was diagnosed when psychological factors were judged to play an etiological role in the development of pain and when organic factors were absent or not severe enough to explain the pain complaint. The diagnostic criteria for somatoform pain disorder in DSM-III-R (14) no longer specified that psychological factors had to play an etiological role. In addition, the primary criterion was changed from pain to a "preoccupation" with pain for at least 6 months. The criterion of absent/insufficient organic findings was removed from DSM-IV for several reasons, including the lack of clear correlation between physical findings and the level of pain and the realization that failure to detect organic factors does not mean that such factors are absent (74). The clinician can now diagnose pain disorder in both acute and chronic pain conditions. Current criteria also allow for the clinician to choose between pain disorder associated with psychological factors and pain disorder associated with both psychological factors and a general medical condition (29).
The current and previous DSM definitions of pain disorder have been widely criticized over the years. In the past, controversy existed as to what constituted an adequate organic finding to explain the pain and at what point pain complaints were considered excessive (19). DSM-IV pain disorder has been criticized for being overly inclusive (75); for requiring subjective value judgment in deciding whether psychological factors have a role in the onset, severity, exacerbation, or maintenance of pain (74, 75); and for being an "unsatisfactory diagnosis of exclusion" in the sense that it is a psychiatric condition for which no distinct psychopathology is specified (50).
The controversy over the current and previous definitions of pain disorder, the change in diagnostic criteria, and the subjectivity required to make these diagnoses have resulted in widely disparate findings regarding prevalence rates in chronic pain populations. For example, Polatin et al. (19) found that 97% of a sample of CLBP patients in an intensive rehabilitation setting had current somatoform pain disorder. In all cases, the disorder developed after the onset of pain associated with an injury. In contrast, only 0.3% of chronic pain patients (73% with CLBP) were diagnosed with psychogenic pain disorder in a study conducted by Fishbain et al. (59). To our knowledge, no studies have addressed the prevalence of pain disorder using the DSM-IV criteria. However, prevalences are expected to be higher in light of the research and theory described in earlier sections of this review, which draw a clear connection between chronic pain and various psychological factors. It appears that pain disorder, by definition, will apply to most or all chronic pain patients.
The nature of the causal relationship between chronic pain and the DSM pain disorders currently receives minimal attention in the research literature. Engel (76) first described the "pain-prone patient," a concept that was later revived (77). Aronoff (46) recently proposed that "pain-prone disorder" be included in the next revision of the DSM. Although these theorists suggest that there are pain-prone personality styles that precede development of chronic pain, most research has demonstrated that there is a general nonspecificity in terms of the relationship between personality/psychological problems and pain (15, 78) and that significant psychopathology typically develops (or recurs) only after months of experiencing disabling pain (15). However, there seems to be a subset of chronic pain patients who demonstrate a consistent tendency to experience and communicate emotional distress as somatic symptoms (79). This syndrome, which may or may not approach the severity of a somatoform disorder, has been called somatization.
Somatization
The construct of somatization has received a great deal of attention in the psychological, medical, and, more specifically, the chronic pain literature. As currently conceptualized, somatization involves the focusing of attention on internal stimuli (80) and the denial of psychological or interpersonal difficulties (81), resulting in an increased tendency to report somatic symptoms, many of which cannot be medically explained (80). Main et al. (82) use the term "somatic distress" to bring attention to the function of somatization as an alternate means of communicating emotional distress. Although somatization is a widespread phenomenon in the chronic pain population (82), perhaps because the reporting of somatic symptoms is viewed as a more appropriate route for seeking help from a physician (83), there seems to be a subset of this population in which this process is amplified (77). For these patients, somatization is conceptualized as a stable trait (diathesis) that becomes activated in response to situations and events the individual finds stressful, such as a painful injury. Although not necessarily meeting the DSM criteria for a somatoform pain disorder, somatization has been found to be associated with increased risk for developing chronic pain and greater health utilization in acute pain patients (84, 85) and poorer treatment outcomes in chronic pain patients (86).
CHRONIC PAIN AND PERSONALITY DISORDERS
According to DSM-IV, "Personality traits are enduring patterns of perceiving, relating to, and thinking about the environment and oneself that are exhibited in a wide range of social and personal contexts. Only when personality traits are inflexible and maladaptive and cause significant functional impairment or subjective distress do they constitute Personality Disorders. The essential feature of a Personality Disorder is an enduring pattern of inner experience and behavior that deviates markedly from the expectations of the individual’s culture and is manifested in at least two of the following areas: cognition, affectivity, interpersonal functioning, or impulse control" (29). Personality disorders (PDs) are thought to develop during childhood or adolescence and must be evident by late adolescence or early adulthood. As long-term patterns, these disorders, by definition, precede the development of pain episodes. DSM-IV recognizes 10 distinct PDs, which are coded on axis II. Research indicates that patients with codiagnoses on axis I and axis II have a poor psychiatric prognosis and higher relapse rate than comparable patients without an axis II diagnosis (87).
Several studies have documented high rates of PDs among chronic pain patients (13, 19, 59, 64, 88). Prevalences ranged from 31% (88) to 81% (64) in these studies. Although no firm statistics are available for overall rates of PDs in the general population, examination of the prevalence of specific PDs indicates that these disorders are more common in the chronic pain population (58, 89). In terms of specific PDs among chronic pain patients, little consistency has been found among different research groups. For example, histrionic (13), dependent (59), paranoid (19), and borderline (88) PDs have been identified as the most common specific axis II disorders in different studies. Discrepancies in patient samples and diagnostic methods (structured vs. unstructured interviews) may be partially responsible for this lack of consistency as well as the lack of consistency in overall rates of PDs between different studies. The overlap between various DSM-IV PD categories (90) may also play a part in these findings.
Additional important findings related to chronic pain and personality disorders have also been discussed in the research literature. Prevalences of PDs were found to be much higher in a sample of CLBP patients (60%) compared with samples of acute LBP patients (24) and acute carpal tunnel syndrome patients (91), further supporting Gatchel’s (2, 18) model of the progression from acute to chronic pain. Gatchel et al. (92) found that the presence of any SCID-diagnosed PD, along with several other psychosocial variables, was predictive of which acute LBP patients had not returned to work 6 months later. The authors noted that no specific type of PD was found to predict chronicity, leading them to suggest that an axis II diagnosis of any type may reflect a general deficit in coping skills that is linked to chronic disability. However, in a follow-up to their study, Gatchel et al. (23) did not find PDs to be predictive of return to work status 1 year later. In patients who already have a chronic condition, inconsistent results are also noted, with some finding no association between PDs and 1-year return-to-work status in a sample of CLBP patients (25), but with others finding an association between a diagnosis of borderline PD and lower 1-year work return rates in a sample of patients with chronic upper extremity disorder (64).
A recent study (27) examined rates of SCID-diagnosed axis I and axis II psychiatric disorders in a sample of CLBP patients before and after (approximately 6 months later) participation in a comprehensive pain rehabilitation program. As expected, rates of several axis I disorders (somatoform pain disorder, MDD) declined with treatment. However, given the chronic nature of PDs and the lack of treatment focus on these disorders in the rehabilitation setting, the authors were surprised to find a decrease in the overall rates of PDs as well as decreases in the rates of specific PDs, including paranoid, obsessive-compulsive, passive-aggressive, and self-defeating PDs. In contrast to these categorical measure results, self-report measures of maladaptive personality traits demonstrated much less change from pre- to posttreatment. The self-report measures used in this study were the Schedule for Nonadaptive and Adaptive Personality (93) and the Minnesota Multiphasic Personality Inventory (94). These measures assess personality pathology from a trait-dimensional perspective in that persons are evaluated on a number of distinct, continuous traits relevant to personality and its dysfunction. The SNAP, for example, assesses three largely independent, higher-order temperament dimensions (negative temperament/neuroticism, positive temperament/extraversion, and disinhibition vs. constraint) and 12 lower-order traits from normative levels through a clinical range of personality pathology. From pre- to posttreatment, none of the higher-order dimensions and only two of the lower-order dimensions changed significantly (aggression and workaholism decreased) despite clinically significant elevations on 12 of the 15 scales at pretreatment. Similar results were obtained with the MMPI, with only a small decrease in scale 0 (social introversion) from pre- to posttreatment. Thus, it seems that self-report trait-dimensional measures of personality functioning are more stable from pre- to posttreatment than categorical PD diagnoses. The SNAP and the MMPI functioned largely as trait measures in this sample, with mean scores and profiles stable over time and a physical rehabilitation program. SCID-derived PD diagnoses failed to show this degree of stability, bringing into question what is being assessed with this structured interview in this population of patients. Vittengl et al. (27) cite regression to the mean or the arbitrary nature of diagnostic cutoff points as possible explanations for these findings. They also suggest that interview-derived PD diagnoses may not represent enduring individual characteristics.
Other explanations for these findings are also possible. Monti et al. (95) indicate that a state of illness can facilitate regressive defenses and accentuate personality traits. Weisberg and Keefe (58, 89) suggest that the findings of the Vittengl et al. study are best explained by a diathesis-stress model of the relationship between chronic pain and PDs. According to these authors, personality patterns that are associated with marginally adaptive coping styles commonly decompensate under the stress of injury, disability, and pain, resulting in the expression of a personality disorder. After treatment that has improved function as a primary goal, stress is likely to decrease. In these cases, individuals who met the criteria for PDs before treatment may still possess the criterion traits, but to a lesser extent than required for a categorical diagnosis. Weisberg and Keefe’s (58, 89) model is intuitively appealing and is also consistent with other diathesis-stress models (2, 18, 37) of the relationship between psychopathology and chronic pain.
CONCLUSIONS
A large body of research has consistently documented elevated rates of psychopathology in patients with chronic pain. Depressive disorders, anxiety disorders, somatoform disorders, substance use disorders, and personality disorders have been identified as the most common diagnostic categories. Although there are unique aspects to the relationship between chronic pain and each specific type of psychopathology, a diathesis-stress model is emerging as the dominant overarching theoretical perspective. In this model, diatheses are conceptualized as preexisting, semidormant characteristics of the individual before the onset of chronic pain that are then activated by the stress of this chronic condition, eventually resulting in diagnosable psychopathology. For example, a premorbid depressive attributional style may predispose an individual to develop a depressive disorder in the context of chronic pain.
The general diathesis-stress model was first proposed by Gatchel (18) as part of his three-stage model describing the development of chronicity in pain patients. More recently, Banks and Kerns (37) applied the diathesis-stress model specifically to the relationship between chronic pain and depression. At this point in time, the diathesis-stress model has received the most support for explaining the development of depressive disorders and personality disorders in the context of chronic pain. Less evidence has accumulated for anxiety disorders, substance use disorders, and somatoform disorders, although future research with these disorders is expected to provide additional support for the model. Conclusive empirical support for such a model will require a prospective research design, given that these diatheses could be validly assessed only before the onset of the chronic pain condition.
Implementing such a prospective research design would be difficult for several reasons. First, it would require assessment of potential diatheses before the onset of chronic pain. Data would have to be collected before the onset of injury or pain (96), or after the onset of acute pain, but before the development of chronic pain (23). In the first case, a very large number of subjects would have to be recruited, because only a small percentage would be expected to be injured and later develop chronic pain. In the second case, involving assessment of diatheses while pain is still acute, fairly large numbers of subjects would still be required, given that only a small percentage would be expected to later develop chronic pain. In addition, there is the possibility that the assessment of diatheses might be confounded by the acute pain experience. A final difficulty in assessing diatheses would be identifying reliable and valid measures of these constructs. If these difficulties could be overcome, and particularly if diatheses could be accurately assessed during the acute pain stage, prophylactic interventions could be used. For example, if negative schemas (53) were identified during an acute pain episode, antidepressant medication or short-term cognitive behavior therapy could be used to prevent the emergence of a full-blown depressive episode.
Finally, it is worth noting that no single model is expected to be able to explain the relationship between chronic pain and psychopathology in all cases. The long search for the answer to the question of what occurs first, the chronic pain or the psychopathology (the so-called "chicken-and-egg" dilemma), has shown that there is no single explanation that can be generalized to all individuals. For example, depression has been shown to be an antecedent of chronic pain, a consequence of chronic pain, and a concomitant biological relative of chronic pain. What does seem certain in almost all cases is that there a prepain vulnerability to a certain type of psychopathology, or a more generalized vulnerability to psychopathology (eg, neuroticism), that may or may not have reached the level of diagnosable psychopathology before the onset of pain. The stress associated with pain then exacerbates the vulnerability or already expressed psychopathology. At the same time, and in a reciprocal manner, psychopathology intensifies the pain experience. Eventually, the dynamic and reinforcing interplay between pain and psychopathology takes on a life of its own, making it impossible to treat either condition independently of the other.
ACKNOWLEDGMENTS
Supported in part by Grants 2R01-MH46452, 2K02-MH01107 and 2R01-DE10713 from the National Institutes of Health.
Received for publication June 12, 2001.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  R. Noyes Jr., S. P. Stuart, and D. B. Watson A Reconceptualization of the Somatoform Disorders Psychosomatics, February 1, 2008; 49(1): 14 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. A. Walsh, E. Jamieson, H. MacMillan, and M. Boyle Child Abuse and Chronic Pain in a Community Survey of Women J Interpers Violence, December 1, 2007; 22(12): 1536 - 1554. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Sharp and B. Keefe Psychiatry in Chronic Pain: A Review and Update Focus, September 1, 2006; 4(4): 573. [Full Text] [PDF]
|
