This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Zautra, A. J. Articles by Smith, B. W. Search for Related Content PubMed PubMed Citation Articles by Zautra, A. J. Articles by Smith, B. W. Related Collections Musculoskeletal Depression Psychophysiology

Depression and Reactivity to Stress in Older Women With Rheumatoid Arthritis and Osteoarthritis

From the Department of Psychology, Arizona State University, Tempe, Arizona.

Address reprint requests to: Alex J. Zautra, P.O. Box 871104, Department of Psychology, Arizona State University, Tempe, AZ 85287-1104. Email: alex.zautra{at}asu.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
OBJECTIVE: The purpose of this study was to examine the role of depressive symptoms in reactivity to stress and pain in older women with rheumatoid arthritis (RA) and osteoarthritis (OA).

METHODS: Participants were 188 older women with RA (N = 87) and OA (N = 101). They were initially assessed for depressive symptoms and interviewed weekly for 12 to 20 weeks regarding interpersonal stress, arthritis pain, and negative affect.

RESULTS: Hierarchical linear modeling (HLM) revealed that depressive symptoms were related to weekly elevations in arthritis pain, negative events, perceived stress, and negative affect for RA respondents and elevations in arthritis pain and negative affect for OA respondents. HLM analyses also indicated that depressive symptoms were related to increased reactivity to perceived stress and arthritis pain in people with RA, but not those with OA.

CONCLUSIONS: Depression may be related to elevations in pain for people with RA and OA and to elevations in stress and increased reactivity to stress and pain for those with RA.

Key Words: depression • arthritis • stress reactivity • pain • women

Abbreviations: RA = rheumatoid arthritis; OA = osteoarthritis; Il-6 = interleukin 6.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
Recent studies on the course of rheumatic disease have provided a strong basis for considering the role of psychosocial factors on the experience of chronic arthritis pain (1–4). Two factors, in particular, have been shown to relate to problems of adaptation among patients with arthritis: stressful life events, especially those of an interpersonal nature, and depression (5–7). In this paper, we examine the role of these factors in reports of pain within two arthritis groups: those with RA, and those with OA.

Both stressful events and responses to stressful events have been related to arthritis pain in studies of RA patients (8). Evidence that stress increases pain sensitivity has been found in a number of studies, and the reasons for suspecting such a relationship are well-founded theoretically (3, 9). Stressful events focus attention on negative affective states and increase awareness and vigilance in response to potential harms (10). Further, some stress responses including feelings of helplessness, catastrophizing, and other coping difficulties that are known to increase psychological distress (11). Reductions in pain have also been reported in therapies that focus on cognitive byproducts of stress, providing further evidence of the importance of these psychological factors to understanding chronic pain (12).

A stress-diathesis model has proven to be a useful paradigm for understanding how stress and individual differences such as depression may play a role in chronic pain (13, 14). In this model, life events and other sources of stress are expected to challenge the adaptive capacity of the organism, but characteristics of the individual, referred to as diathesis factors, determine the degree of susceptibility to stressful events.

Depressive symptoms have been associated with chronic pain in a number of studies. These symptoms are associated with a number of processes thought to increase pain, including cognitive distortions (15), less positive reinforcement (16), helplessness (7), and a relative lack of positive affect (17). One of the mechanisms by which pain may increase is through a failure to mount an adequate psychological defense to life stresses. Indeed, heightened negative affect in response to stress has been shown to increase pain sensitivity (2). Depression may also influence pain through over-attention to pain and other negative affects, and relative inattention to situations that provide the means for positive affective engagement (18).

When examining the contribution of individual differences in stress reactivity, there are also several links to consider in the chain of events following stressful events (19). These linkages include the occurrence of the stressor itself, appraisals of the stressfulness of the event as revealed in perceptions of stress, followed by a rise in pain sensitivity. Some patients with chronic pain have more stress in their lives. Further, the frequency of some forms of stress may be a consequence of depression and/or other individual differences (20). The value of distinguishing between stressor and the stress response is particularly evident in the study of interpersonal stress. Interpersonal stressors are known to significantly affect the arthritis patients’ sense of well-being as well as their reports of pain (21, 22). Depressive symptoms may play a role in generating interpersonal stress, and/or increasing vulnerability to interpersonal tensions that arise independently of the person’s psychological status.

In addition to psychosocial mechanisms of vulnerability to chronic pain, biological factors can make an important difference in susceptibility to stress. Among the many painful conditions subsumed under the umbrella of arthritis, two of the most prevalent forms, RA and OA, are fundamentally different in their pathophysiology (23, 24). RA is a systemic autoimmune disease characterized by chronic inflammation in joints on both side of the body, organized symmetrically. The age of onset for RA is typically between 35 and 45, and women are afflicted with the disease three times more often than men. OA, on the other hand, is characterized by local inflammation in single joints without involvement in the symmetrical joint on the other side of the body. OA is not due to a systemic autoimmune process unlike RA. In contrast, OA is typically referred to as a "wear and tear" disease, with joint damage due to use, and/or misuse over time. The age of onset of OA is 50 to 60 for most, and is prevalent in both older men and women.

The pain from both OA and RA may be influenced by stress processes through psychological mechanisms such as the ones noted above (25, 26). However, because RA is an autoimmune disease, there is reason to suspect that psychosocial factors would influence disease course through influences on the immune system as well (27). Both stress and depression have been shown to affect immune processes associated with RA (28–30). A blunted cortisol response has been proposed as a potential cause of the body’s failure to downregulate immune responses in RA (31, 32). Cortisol increases in response to stress ordinarily, and has been associated with depressive symptoms, as well as autonomic arousal following stress (33, 34). Such influences would not be found for OA since the mechanisms of inflammation are primarily in the joint itself, and not due to autoimmune dysfunction.

The approach we took in this study was to examine the relationships between stressors, stress responses, and pain for patients with a diagnosis of either RA or OA. We expected to find some evidence of a relationship between responses to stress and pain for both OA and RA patients as predicted by psychosocial models of the role of stress in furthering attention to pain, and reducing the capacity to cope successfully. Further, we examined the role of depressive symptoms in predicting pain levels overall, and in amplifying pain signals during times of stress for both groups. We hypothesized differences between these two groups in the relationships between stress and pain. Because of the systemic nature of RA, we predicted that RA patients would show a greater reactivity to stress if they were depressed, and that reactivity we expected to translate into a greater elevation in pain following stressful events for RA patients, when compared with those with OA.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
Participants
The sample consisted of 188 women who had rheumatoid arthritis (RA; N = 87) or osteoarthritis (OA; N = 101). Only postmenopausal women were studied here to reduce within and between group heterogeneity in the hormonal profiles that might underlie stress-related inflammatory processes. The participants were recruited in a variety of ways including newspaper ads, mailings to Arthritis Foundation members, and through rheumatology clinics. All participants had their diagnosis (RA or OA) confirmed by a physician and were paid a total of $100 ($50 after the initial interview and $50 after the final interview).

Participants were between the ages of 42 and 76 (M = 63.8, SD = 7.2). The majority were married (55.6%) while fewer were divorced (23.8%), widowed (16.4%), or never married (3.7%). The majority were Caucasian (95.2%) while a small proportion were Hispanic (2.7%), African-American (1.1%), or Native-American (1.1%). The mean income range was between $21,000 and $25,000. Twenty-nine percent had a yearly income of less than $20,000, 42% had an income between $20,000 and $40,000, and 29% had an income of more than $40,000. Ninety-four percent completed high school and 29% completed four years of college. Twenty-six percent were employed.

There was a slight but significant difference between the RA and OA respondents in age. The mean age was 62.3 (SD = 7.31) for people with RA and 65.1 (SD = 6.92) for people with OA (t 187 = 2.75, p < .01). There were no notable differences between the groups on marital status, ethnic background, income, or percent employed (p values > .30).

Procedure
The data used for this study came from an initial questionnaire mailed to participants and from weekly telephone interviews conducted by trained research assistants. Depressive symptoms were measured in the initial questionnaire. Negative and positive interpersonal events, perceived interpersonal stress, arthritis pain, and negative affect were assessed during the weekly telephone interviews.

Participants were interviewed weekly for between 12 and 20 weeks. Participation was discontinued after 12 weeks if they met the criteria for a stressful week and an arthritis flare. If they had not met these criteria by 12 weeks, they continued in the study for up to 20 weeks until they met criteria. Twelve weeks was chosen as the minimum length of time that would provide sufficient weekly data points to evaluate the relationship between interpersonal stress and disease activity. However, those with low levels of disease activity and interpersonal stress were kept in the study longer to increase the chance of obtaining variability on these factors. The mean number of weeks the participants were in the study was 17.1. The total number of weeks in the study for all participants was 3209 (188 participants x 17.1 weeks). The total number of weeks of data the participants provided was 2980 (3209 total weeks minus 229 weeks of missing data).

Measures
Arthritis pain.
The average level of arthritis pain was assessed each week with the following instruction: "Please choose a number between 0 and 100 that best describes the average level of pain you have experienced over the past week due to your arthritis. A zero [0] would mean ’no pain’ and a one hundred [100] would mean ’pain as bad as it could be.’" The week-to-week test-retest correlation was 0.87 for RA respondents and 0.52 for those with OA.

Depressive symptoms.
Depressive symptoms (Fig. 2) were measured at the beginning of the study. The measure for depressive symptoms consisted of nine items taken from the Mental Health Inventory (35). The items included questions such as "How often have you felt like crying?" and "How much of the time have you felt downhearted and blue?" All nine items are included in the Appendix. Participants were asked to respond to each of these items in terms of the previous week. The responses were scored on a six-point scale for all items except for the last which was scored on a five-point scale. Cronbach’s was 0.93 for RA respondents and 0.91 for those with OA.

Negative affect.
Negative affect was measured using ten items taken from the Positive and Negative Affect Schedule (36). Participants were asked to indicate the extent to which they had experienced negative affects during the previous week. Participants were asked to respond to a five-point scale where 1 = "very slightly or not at all" to 5 = "extremely." The negative affect items were "distressed," "upset," "guilty," "scared," "hostile," "irritable," "ashamed," "nervous," "jittery," and "afraid." Negative affect was measured during each week of the study. The week-to-week test-retest correlation was 0.66 for RA respondents and 0.54 for OA respondents. Cronbach’s was 0.86 for RA respondents and 0.82 for those with OA.

Although both depressive symptoms and negative affect can both be conceived of as aspects of psychological distress, they were measured and considered separately in our study. Whereas negative affect is a measure of current state, defined in terms of negative emotions, the depressive symptoms variable is measured here as a more stable trait, and is defined more broadly in terms of emotion (presence of negative affect and absence of positive affect), cognition (helplessness, hopelessness, rumination), social relations (loneliness, isolation, withdrawal) and physical symptoms (lack of energy). Indeed, the correlation between initial depressive symptoms and initial negative affect in our study was 0.34 suggesting that there is a large degree of nonoverlapping variance (88%). Thus the depressive symptoms variable was of interest in this study because of its comprehensive nature as a stable indicator of clinical depression. Negative affect was assessed weekly as a state variable that was expected to covary with weekly fluctuations in interpersonal stress and arthritis pain.

Negative and positive interpersonal events.
Negative and positive interpersonal events were assessed each week by the Inventory of Small Life Events (37). All items on this list were written to have a discrete beginning and end. They refer to observable changes in everyday activities rather than internal states. Participants were asked to report the number of times each event happened during the past week. Negative and positive events were measured in relation to four kinds of relationships: spouse or significant others, family members, friends, and coworkers.

Examples of the negative events assessed include having arguments, being criticized, meeting unfriendly or rude people, and pressure to work harder or faster. Examples of the positive events assessed include receiving expressions of love and affection, receiving gifts, going out with friends, having a party, and being praised.

The scores for negative events and positive events did not always contain the items for all four categories of relationship because not all participants had significant others or were employed. The mean score for all the items that applied to each participant was taken as the total score. For example, if the participant was not employed but was married and had family members and friends, then only the significant other, family member, and friend items were used to calculate the total mean score. Also, a maximum of one event per day for each specific event was allowed and the square root of total scores was used in the analyses. These procedures were used in order to minimize the influence of the extreme scores and kurtosis that is common with event measures. The week-to-week test-retest correlation for negative events was 0.62 for people with RA and 0.68 for those with OA. The test-retest correlation for positive events was 0.70 for RA respondents and 0.77 for those with OA. Cronbach’s were not computed for negative and positive events. Since the event items were written to be independent of one another, we did not see internal consistency as an appropriate test of reliability for these measures.

Perceived interpersonal stress.
Perceived interpersonal stress was measured in terms of the categories of spouse or significant other, family members, friends, and coworkers. Participants were asked "Overall, how stressful were your relations with (category of relationship) this past week?" The responses were scored on a four-point scale where 0 = "not stressful at all" to 3 = "extremely stressful." As with the negative and positive interpersonal event measures, the total score was created by calculating the mean for the items from the categories that applied to the participant. If items from one or two categories were missing (eg, because they did not have a significant other or were not employed), then the mean was taken for the remaining items. Perceived interpersonal stress was measured weekly. Cronbach’s was not computed because the items were designed to measure nonoverlapping domains of interpersonal relationships. The week-to-week test-retest correlation was 0.55 for RA respondents and 0.82 for those with OA. "Perceived interpersonal stress" will be abbreviated to "perceived stress" for the remainder of this article.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
Initial analyses provided the descriptive information and intercorrelations among depressive symptoms and the variables measured each week. For these analyses, the weekly variables were aggregated for each participant. Table 1 displays the results of these analyses both for people with RA and those with OA. There were no notable differences between the two groups on the mean for depressive symptoms or any of the weekly variables. Next, we inspected the intercorrelations among depressive symptoms and each of the weekly variables for both RA and OA groups. Instead of computing the correlations at any one point in time, we used data from all weeks of the study in order to find the best approximation of the correlations involving the variables measured each week. Since observations were not independent, significance tests for N = 188 (number of participants) rather than for N = 2980 (number of observations) were used. Table 2 displays the results for the RA and OA groups. As expected, for both groups, depressive symptoms were related to arthritis pain and negative affect. In addition, negative affect, negative events, and perceived stress were all related for both groups.

Hierarchical linear modeling (HLM) was used to address each of these questions because they involve weekly data which has a hierarchical structure with 12 to 20 observations nested within each of the 188 participants. All HLM analyses were conducted using the SAS PROC MIXED software. An explanation of the hierarchical models used in this study can be found in Bryk and Raudenbush (39) and Singer (40). Finally, RA respondents and those with OA were analyzed separately because of the difference in nature of the diseases (see the Introduction).

Perceived Stress and Arthritis Pain
First, we wanted to know whether perceived stress was related to elevations in arthritis pain during the same week. An HLM equation was specified in which perceived stress was the predictor and arthritis pain was the outcome. The equation was specified as follows:

Age was controlled because there was a meaningful difference in age between the RA and OA samples. No other demographic variables were controlled because there were no other notable difference between groups (p values > 30). The week number in the study was also used as a control variable because the mean level of the weekly variables tended to decrease as the number of weeks in the study increased. This occurred because participants were likely to be in the study longer if they had lower mean scores due to the criteria for study completion.

In addition, a first order autoregressive variance-covariance matrix was used to model the within-subjects covariance on the dependent variable. An autoregressive matrix was chosen because it allows for the examination of week-to-week changes in the dependent variable. The autoregressive matrix does this by removing variance in the dependent variable predicted from previous weeks’ scores. The first order autoregressive function expects a linear relationship between adjacent scores and an exponential decay in that relationship the further apart in time the measures are taken. The result is that the dependent variable scores are residualized of their previous scores.

The results indicated that, for people with RA, perceived stress was related to elevations in arthritis pain (B = 4.307, p < .001). For people with OA, perceived stress was related to elevations in arthritis pain (B = 1.596, p < .05, one-tailed). An additional analysis was conducted to determine whether there was a significant difference between the RA and OA respondents in the strength of the relationship between perceived stress and arthritis pain. Diagnosis was added as a dummy variable and the diagnosis by perceived stress was marginally notable (B = 2.568, p < .05, one-tailed) indicating that perceived stress may have a stronger relationship with arthritis pain for people with RA.

Direct Effects of Depressive Symptoms
Next, the relationship between initial depressive symptoms and weekly changes in arthritis pain and each of the other weekly variables was examined. HLM equations were specified in which depressive symptoms was the predictor and each weekly variable was the outcome. The equations that were specified are as follows:

For the reasons given above, age and the week number in the study were used as control variables and an autoregressive covariance matrix was used to model the within-subjects variance on the dependent variable.

Depressive symptoms were related to changes in all five of the weekly variables for people with RA and three of the weekly variables for those with OA. Table 3 displays the results for each group. For the RA respondents, depressive symptoms predicted weekly elevations in arthritis pain, negative affect, negative events, perceived stress, and weekly decreases in positive events. For the OA respondents, depressive symptoms predicted weekly elevations in arthritis pain and negative affect and weekly decreases in positive events. An additional analysis using diagnosis as a dummy variable and examining the diagnosis by depressive symptoms interaction were conducted to examine differences between the RA and OA respondents. There was a meaningful difference in the strength of the relationship between depressive symptoms and negative affect (B = 0.150, p < .05) showing a stronger relationship for people with RA. None of the other differences were significant.

For the reasons given above, age and the week number in the study were used as control variables and an autoregressive covariance matrix was used to model the within-subjects variance on the dependent variable.

First, we predicted elevations in perceived stress from depressive symptoms, negative events, and the depressive symptoms by negative events interaction. For people with RA, elevations in perceived stress were predicted by negative events (B = 0.070, p < .001) and marginally by depressive symptoms (B = 0.061, p < .10) but not by the depressive symptoms by negative events interaction. For those with OA, elevations in perceived stress were predicted by negative events (B = 0.076, p < .001), but not by depressive symptoms or the depressive symptoms by negative events interaction.

Second, we predicted elevations in arthritis pain from depressive symptoms, perceived stress, and the depression by perceived stress interaction. The results for the RA and OA respondents are displayed in Table 4. For people with RA, depressive symptoms, perceived stress, and the depressive symptoms by perceived stress interaction were all meaningful predictors of elevations in arthritis pain. The depressive symptoms by perceived stress interaction was positive indicating that the relationship between perceived stress and arthritis pain was stronger for those higher in depressive symptoms. For those with OA, depressive symptoms was a consequential predictor and perceived stress was a marginal predictor of arthritis pain. The interaction was not a meaningful predictor for the OA respondents.

Third, we wanted to determine whether depression might result in a continuation of the cycle of stress and pain by increasing negative events during the following week. We predicted elevations in negative events during the following week from depressive symptoms, arthritis pain, and the depressive symptoms by arthritis pain interaction. For people with RA, depressive symptoms and the depressive symptoms by arthritis pain interaction were notable predictors of elevations in negative events during the following week (B = 0.893, p < .05 and B = 0.012, p < .05, respectively). The depressive symptoms by arthritis pain interaction was positive indicating that the relationship between arthritis pain and negative events the following week was stronger for those higher in depression. For people with OA, depressive symptoms, arthritis pain, and the depressive symptoms by arthritis pain interaction were each not important predictors of negative events during the following week.

An additional analysis was conducted to determine whether the interaction was extensively different for the RA and OA samples. Diagnosis was added as a dummy variable and the three-way interaction (diagnosis by depressive symptoms by arthritis pain) was not notable (B = 0.007, ns). Thus, depression did not play a considerably stronger role for people with RA in increasing the strength of the relationship between arthritis pain and future negative events.

Since the perception or actual occurrence of negative events may be linked to negative affect, we suspected that the depressive symptoms by arthritis pain interaction would be related to increased negative affect during the following week for people with RA. Indeed, the depressive symptoms by arthritis pain interaction was a notable predictor of elevations in negative affect during the following week for those with RA (B = 0.003, p < .01), but not for those with OA. An analysis using diagnosis as a dummy variable and examining the triple interaction (diagnosis by depressive symptoms by arthritis pain) did not reveal a important difference between the RA and OA respondents in the role depressive symptoms played in the association between arthritis pain and future negative affect.

The results of the moderational analyses from this study were incorporated into our model of the stress-pain process in arthritis patients. The revised model is presented in Fig 1 for both people with RA and those with OA. The previous model was that negative events was related to the perception of stress which, in turn, was related to increased reports of arthritis pain. These two relationships were confirmed in the RA and the OA samples. However, depressive symptoms appeared to play a more important role for the RA respondents as compared with those with OA. For people with RA, depressive symptoms had direct effects in increasing negative events, perceived stress, and arthritis pain and increased the strength of the relationship between perceived stress and arthritis pain and between arthritis pain and future negative events. For those with OA, there was only a direct effect of depressive symptoms in increasing arthritis pain.

View larger version (0K): [in this window] [in a new window]   Fig. 1. The effects of depression on the stress-pain process in rheumatoid arthritis and osteoarthritis patients. All lines drawn are notable effects, positive in sign. Tests of the relationship between the variables measured during the "Current Week" were cross-sectional. In contrast, tests of arthritis pain during the "Current Week" leading to negative events the "Following Week" were prospective.

The results are displayed in Table 5. For people with RA, positive events, perceived stress, and the positive events by perceived stress interaction were substantially related to arthritis pain. Positive events predicted less pain, perceived stress predicted more pain, and positive events buffered the relationship between perceived stress and arthritis pain. Positive events and the positive events by perceived stress interaction did not predict arthritis pain in those with OA. An analysis using diagnosis as a dummy variable and examining the triple interaction (diagnosis by positive events by perceived stress) was used to compare the RA and OA samples. This analysis revealed that positive events played a importantly stronger role for people with RA in buffering the relationship between perceived stress and arthritis pain (B = 1.402, p < .01).²

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
Both RA and OA participants reported greater pain during weeks in which they reported greater interpersonal stress. Overall, a pattern emerged from these data to suggest that weeks when respondents reported more stressful interpersonal events they also reported greater perceived stress in interpersonal domains. The perceived stress was associated with greater pain among those with arthritis. Such findings were important to document within a study such as this one that used more careful methods of measurement, and a larger sample of arthritis participants than has been employed in most prior research on this topic.

Our past work also led us to make additional predictions that were not as straightforward. We reasoned that there would be important individual differences in the vulnerability to stress among those in chronic pain. Other researchers have argued that reactivity to stressful events is likely to be greater among depressed participants. We expected to observe those effects primarily among participants with active rheumatoid arthritis, not those with osteoarthritis, because of fundamental differences in pathophysiology in the two diseases. Consistent with these ideas we tested the effects of depression on the stress-pain relationships, and did so separately for the two arthritis groups, in order to examine whether RA participants revealed a greater reactivity to interpersonal stress than the OA sample.

The pattern of results were consistent with the hypotheses that depression would be related to greater stress reactivity in the RA vs. the OA group. Although RA participants were not more depressed than OA participants, their depression had greater influence on the relationships between stress and pain for RA participants compared with their OA counterparts. These effects were not due to differences in level of stress or pain between groups. Depression was a risk factor for pain in both OA and RA samples, but only in RA was depression a risk factor for stress reactivity.

The OA group did not show better adaptation overall. They had the same level of depression and also the same level of pain as those with RA. The mechanisms by which stress and emotions influence health outcomes for these two groups appear to be different at some basic level, however. Indeed, the findings suggest that the relationships between psychosocial factors like interpersonal stress and depression depend in part on the biologic condition under study (1). Our first guess about these data are that the effects on the RA sample are the consequence of pathogenic inflammatory processes that can be influenced directly, as well as indirectly, by stress and distress. The OA participants, without the same proinflammatory pathologic makeup, show less reactivity.

What are the psychological factors that provoke stress reactivity in the RA? Depressive symptoms represent a broad set of emotions and cognitions, including feelings of helplessness, despair and other negative affects, loss of incentive, and an absence of positive affect as well (41). Those with higher initial scores on depressive symptoms reported more negative events, higher perceived stress, more weekly elevations in negative affect, and more arthritis pain. We also probed the role of a relative absence of positive interpersonal experiences as one potential aspect of depression that might also be responsible for the effects observed. Positive interpersonal events did play a role in influencing the effects of perceived stress on pain for the RA group in much the same way that depressive symptoms did. In a similar vein, recent studies of the clinical course of cardiac patients found that negative affectivity when coupled with social inhibition was highly predictive of subsequent cardiac events (42).

What physiological process might be responsible? There are a number of possible candidates among those elements of the immune system responsible for a pro-inflammatory autoimmune response. One likely agent is Il-6, a neuropeptide known to be elevated in acute depression. Il-6 stimulates T-helper cells, and has been observed to be elevated among patients with RA as well as depression. It also seems likely that hormonal responses, such as a blunted cortisol response to stress, could tip the scales in favor of a proinflammatory response following stress in the RA person who was also depressed (43). No research has been conducted as yet with RA samples on the contribution of homeostatic mechanisms such as oxytocin, a hormone that increases with positive interpersonal relations and speeds physiological recovery from stressful events (44).

There also are psychosocial forces at work that favor greater reactivity for RA, independent of their disease. First, RA is a source of stress in itself. The disease is less predictable in its course, more crippling, and potentially more devastating to quality of life and social relations than OA. Since multiple joints are involved for RA, more activity limitation is evident among those with that disease compared with OA. In this study, we anticipated that difference and selected OA and RA participants based on a minimum qualification of some limitation in functioning and pain. However, RA participants had their arthritis longer, consistent with the earlier age of onset for RA vs. OA. All these factors could add up to increase stress reactivity, regardless of differences in disease pathophysiology. In this study, however, the two arthritis groups were comparable on psychosocial factors that might reflect such differences in the stressfulness of the chronic illness. For instance, the groups did not differ in initial levels of depression, weekly levels of negative affect, number of interpersonal stressors, or perceived stress.

It is important to note that the lack of difference between groups on the study variables is a result that cannot be generalized to populations of OA and RA patients. Epidemiological methods were not followed in the sample selection to obtain representative samples of these groups. Instead, the OA sample was drawn as a comparison group, and thus was engineered to be more like RA samples than is typically found in the population. This is most evident in the selection of only postmenopausal women in both groups to remove the influence of gender and hormonal fluctuations in the comparisons made between groups and over time. The increase in homogeneity has a clear cost in loss of generalizability. Our findings are limited in their applicability to older adult women in moderate or greater pain from arthritis.

Another limitation of our study was our reliance on depressive symptoms recorded at the baseline interview as the only indicator of depression. This assessment of depression had the advantage of providing a prospective test of how weekly fluctuations in stress and pain could be predicted by initial depression levels. In addition, our psychometric examinations of this measure have revealed that it provides highly stable estimates of depressive symptoms for older adult populations (45). However, depressive symptoms can also vary over time, and measures taken closer to the weekly outcomes of pain might have revealed a stronger pattern of relationships.

While the findings provide support for the role of the disease itself, we suggest that some caution is in order in accepting this interpretation of the data. The group assignments are not random, and differences in selection methods for RA vs. OA samples may have contributed in hidden ways to the results obtained. Many willing RA volunteers did not pass the careful screening because of use of immune-modifying drugs or high dosage levels of prednisone. OA volunteers were screened out of the study if they had a history of autoimmune disease, but not RA recruits. It is impossible to know what effects these many subtle influences on sample selection may have had on the composition of the final sample, and their stress reactivity.

Implications
Engel’s (46) biopsychosocial model has been a frequent rallying point for those within health psychology in their call for greater interdisciplinary attention to disease processes. This research was a product of this clarion call. Up to now, the model has been used primarily to promote the utility of psychosocial processes in understanding the onset and course of disease. Our findings add further support to the proposition that psychosocial processes affect illness outcomes. However, our results take us full circle in the quest for interdisciplinary work. Here we find that the underlying pathophysiology appears to determine the degree of influence of psychosocial factors such as depression. This is a point that is often lost in the debate between biomedical communities and advocates for bio-behavioral mechanisms. We hope that findings of our study will serve as a reminder that not only are biological processes in the body, psychosocial processes may be found to influence health and illness from within as well. Thus the influence of psychological factors such as depression depends on the state of the organism, and that includes biological state as well as his or her behavioral, cognitive, and affective state.

	APPENDIX

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
Depressive Symptoms Items

1. How often did you feel that you had nothing to look forward to?
   
2. How often have you felt like crying?
   
3. How often did you feel that nothing turned out for you the way you wanted it to?
   
4. How often have you felt so down in the dumps that nothing could cheer you up?
   
5. How much of the time have you felt lonely?
   
6. How much of the time have you felt downhearted and blue?
   
7. How much of the time have you been moody and brooded about things?
   
8. How much of the time have you been in low or very low spirits?
   
9. In the past week, did you feel depressed?
   

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
This research is supported by Grant 5 R01 AR41687-05 from the National Institute of Arthritis and Muscular-Skeletal and Skin Diseases, and a grant from The Arthritis Foundation.

	NOTES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 
¹ Each week respondents reported on joint tenderness and activity limitation in addition to pain. A score indicative of disease activity was constructed from these ratings based on prior research on this measure (22). The HLM analyses reported for pain were also conducted for this measure. With the exception of findings for stressful events the following week, all findings notable for pain were also meaningful using this other measure, including the depression by diagnosis interactions.

² To further explore the combined effect of being depressed and having RA, we conducted a subgroup analysis examining the relationship between diagnosis, depression, and a clinician’s rating of disease activity. Data were available on clinician’s global ratings of disease activity taken at a baseline of low stress and disease activity, during a time of high stress, and one week later for 32 RA and 40 OA respondents. These data were analyzed comparing RA and OA respondents who were high vs. low in depression. There were no differences in global ratings as a function of diagnosis or a diagnosis by depression interaction (p > .40). Depression was marginally related to higher disease activity (F(1,68) = 3.16, p < .09). Two interaction effects were observed over time. RA respondents showed elevations in disease activity that lasted into the week following stress, but people with OA returned to baseline levels (F(2,67) = 7.42, p < .03). Second, depressed respondents showed longer lasting elevations in disease activity as well, (F(2,67) = 4.40, p < .02). Analyses focused solely on baseline to the post stress week clinician ratings revealed notable interactions for depression and diagnosis on disease activity with p values < .009. Consistent with the weekly data, these results showed greater reactivity to stress for people with RA and depressed participants.

Received for publication November 30, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX NOTES ACKNOWLEDGMENTS REFERENCES

 

1. Huyser B, Parker J. Stress and rheumatoid arthritis: an integrative review. Arthritis Care and Research 1998; 11: 135–45.
2. Melzack, R. Pain. Past, present and future. Can J Exp Psychol 1993; 47: 615–29.[Medline]
3. Schwartz L, Slater MA, Birchler GR. Interpersonal stress and pain behaviors in patients with chronic pain. J Consult Clin Psychol 1994; 62: 861–4.[Medline]
4. Bijlsma JWJ, Huiskes, CJAE, Kraaimaat FW, Vanderveen MJ, Huber-Bruning O. Relation between patients’ own health assessment and clinical and laboratory findings in rheumatoid arthritis. J Rheumatol 1991; 18: 650–3.[Medline]
5. Fifield J, Tennen H, Reisine S, McQuillan J. Depression and the long-term risk of pain, fatigue, and disability in patients with Rheumatoid Arthritis. Arthritis Rheum 1998; 41: 1851–7.[Medline]
6. Manne S, Zautra AJ. The effects of spouse critical remarks on psychological adjustment in rheumatoid arthritis. J Per Soc Psychol 1989; 56: 608–17.
7. Smith TW, Peck JR, Ward JR. Helplessness and depression in rheumatoid arthritis. Health Psychol 1990; 4: 377–89.
8. Zautra AJ, Hamilton NA, Potter P, Smith B. Field research on the relationship between stress and disease activity in rheumatoid arthritis. Annals New York Academy of Sciences 1999; 879: 397–412.
9. Affleck G, Tennen H, Urrows S, Higgins P. Person and contextual features of daily stress reactivity. J Per Soc Psychol 1994; 66: 329–40.
10. Ursin H, Olff M. The stress response. In: Stanford, SC, Salmon, R, editors. Stress: from synapse to syndrome. San Diego, CA: Academic Press; 1993. p. 3–22.
11. Brown GK, Nicassio PM, Wallston K. Pain coping strategies and depression in rheumatoid arthritis. J Consult Clin Psychol 1989; 57: 652–7.[Medline]
12. Keefe FJ, Caldwell DS. Cognitive behavior control of arthritis pain. Advances in Rheumatology 1997; 81: 277–90.
13. Banks SM, Kerns RD. Explaining the high rates of depression in chronic pain: a diathesis-stress framework. Psychol Bull 1996; 119: 95–110.
14. Monroe SM, Simons AD. Diathesis-stress theories in the context of life stress research: implications for the depressive disorders. Psychol Bull 1991; 110: 406–25.[Medline]
15. Beck AT. The diagnosis and management of depression. Philadelphia, University of Pennsylvania Press 1973.
16. Lewinsohn P. The behavioral study and treatment of depression. In: Hersen, M, Eisler, R, Miller, P, editors. Progress in Behavior Modification. New York: Academic Press; 1975. p. 16–65.
17. Clark LA, Watson D. Tripartite model of anxiety and depression: psychometric evidence and taxonomic implications. J Abnorm Psychol 1991; 100: 316–36.[Medline]
18. Zautra AJ, Hamilton N, Burke HM. Comparison of stress responses in women with two types of chronic pain: fibromyalgia and osteoarthritis. Cognitive Therapy and Research 1999; 23: 209–30.
19. Lennon MC, Dohrenwend BP, Zautra AJ, Marbach JJ. Coping and adaptation to facial pain in contrast to other stressful life events. J Per Soc Psychol 1990; 59: 1040–50.
20. Bolger N, Zuckerman A. A framework for studying personality in the stress process. J Per and Soc Psychol 1995; 69: 890–902.
21. Jamison RN, Virts KL. The influence of family support on chronic pain. Behav Res Ther 1990; 28: 283–7.[Medline]
22. Zautra AJ, Hoffman J, Potter PT, Matt KS, Yocum D, Castro L. Examination of changes in interpersonal stress as a factor in disease exacerbations among women with rheumatoid arthritis. Ann Behav Med 1997; 19: 279–86.[Medline]
23. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries, JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Harvinder SL, Medsger TA, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988; 31: 315–24.[Medline]
24. Yelin E. Arthritis: the cumulative impact of a common chronic condition. Arthritis Rheum 1992; 35: 489–97.[Medline]
25. Watson D, Pennebaker JW. Health complaints, stress, and distress: exploring the central role of negative affectivity. Psychol Rev; 96: 234–54.
26. Wells KB, Golding JM, Burnam MA. Chronic medical conditions in a sample of the general population with anxiety affective and substance use disorders. Amer J Psychiatry 1989; 146: 1440–6.[Abstract/Free Full Text]
27. McEwen BS. Protective and damaging effects of stress mediators. Seminars in Medicine of the Beth Israel Deaconess Medical Center 1998; 338: 171–9.
28. Maes M, Meltzer HY, Bosmans E, Bergmans R, Vandoolaeghe E, Ranjan R Desnyder Increased plasma concentrations of interleukin-6, soluble interleukin-6, soluble interleukin-2, and transferrin receptor in major depression. Journal of Affective Diseases 1995; 34: 301–9.
29. Zautra AJ, Burleson MH, Matt KS, Yocum D, Castro L. Interpersonal stress, depression, hormones, and disease activity in rheumatoid arthritis and osteoarthritis patients. Health Psychol 1994; 13: 139–48.[Medline]
30. Papanicolaou DA, Wilder RL, Manolagas SC, Chrousos GP. The pathophysiologic roles of interleukin-6 in human disease. Ann Intern Med 1998; 128: 127–37.[Abstract/Free Full Text]
31. Sternberg EM, Licino J. Overview of neuroimmune stress interactions implications for susceptibility to inflammatory disease. In: Chrousos GP, McCarty, R, Pacak, K, Cizza, G, Sternberg, E, Gold, PW, Kvetnansky, R, editors. Stress basic mechanisms and clinical implications. New York: Annals of the New York Academy of Science 1995, p. 364–71.
32. Chrousos, GP. The Hypothalamic-Pituitary-Adrenal Axis and immune-mediated inflammation. N Engl J Med 1995; 332: 1351–62.[Free Full Text]
33. Chrousos GP. Stressors, stress, and neurendocrine integration of the adaptive response. In: Csermely, P, editor. Stress of life. New York: The New York Academy of Sciences; 1998. p. 311–20.
34. Pike JL, Smith TL, Hauger RL, Nicassio PM, Patterson TL, McClintick J, Costlow C, Irwin MR. Chronic life stress alters sympathetic, neuroendocrine, and immune responsivity to an acute psychological stressor in humans. Psychosom Med 1997; 59: 447–57.[Abstract/Free Full Text]
35. Veit CT, Ware JE, Jr. The structure of psychological distress and well-being in general populations. J Consult Clin Psychol 1983; 51: 730–42.[Medline]
36. Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect. J Per Soc Psychol 1988; 54: 1063–70.
37. Zautra AJ, Guarnaccia CA, Dohrenwend BP. Measuring small life events. Am J of Community Psychol 1986; 14: 629–55.
38. Affleck G, Tennen H, Urrows S, Higgins, P. Individual differences in the day-to-day experience of chronic pain: a prospective daily study of rheumatoid arthritis patients. Health Psychol 1991; 10: 419–26.[Medline]
39. Bryk AS, Raudenbush SW. Hierarchical linear models: applications and data analysis methods. Newbury Park, CA: Sage; 1992.
40. Singer JD. Using SAS PROC MIXED to fit multilevel models, hierarchical models, and individual growth models. Journal of Educational and Behavioral Statistics 1998; 24: 323-55.
41. Brown, TA, Chorpita BF, Barlow DH. Structural relationships among dimensions of DSM-IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. J Abnorm Psychol 1998; 107: 179–92.[Medline]
42. Denollet J, Brutsaert DL. Personality, disease severity, and the risk of long-term cardiac events in patients with decreased ejection fraction after myocardial infarction. Circulation 1998; 97: 167–73.[Abstract/Free Full Text]
43. Walker JG, Littlejohn GO, McMurray NE, Cutolo M. Stress system response and rheumatoid arthritis. Br J Rheumatol 1999; 38: 1050–7.[Abstract/Free Full Text]
44. Uvas-Moberg K. Oxytocin may mediate the benefits of positive social interaction and emotions. Psychoneuroimmunology 1998; 23: 819–35.
45. Guarnaccia C, Zautra AJ. Older adult mental health factor structure stability: self-reports and confidant reports. Journal of Mental Health and Aging 1998; 4: 307–22.
46. Engel GL. The need for a new medical model: a challenge for biomedicine. Science 1997; 196: 129–36.

This article has been cited by other articles:

				T. P. Stitik, G. Gazzillo, and P. M. Foye State of the Art Review: Osteoarthritis and Therapeutic Exercise American Journal of Lifestyle Medicine, October 1, 2007; 1(5): 360 - 366. [Abstract] [PDF]

				P. E. Stang, N. A. Brandenburg, M. C. Lane, K. R. Merikangas, M. R. Von Korff, and R. C. Kessler Mental and Physical Comorbid Conditions and Days in Role Among Persons With Arthritis Psychosom Med, January 1, 2006; 68(1): 152 - 158. [Abstract] [Full Text] [PDF]

				A. J. Zautra, R. Fasman, J. W. Reich, P. Harakas, L. M. Johnson, M. E. Olmsted, and M. C. Davis Fibromyalgia: Evidence for Deficits in Positive Affect Regulation Psychosom Med, January 1, 2005; 67(1): 147 - 155. [Abstract] [Full Text] [PDF]

				T. Covic, B. Adamson, D. Spencer, and G. Howe A biopsychosocial model of pain and depression in rheumatoid arthritis: a 12-month longitudinal study Rheumatology, November 1, 2003; 42(11): 1287 - 1294. [Abstract] [Full Text] [PDF]

				A. L. Hassett and L. H. Sigal Unforeseen Consequences of Terrorism: Medically Unexplained Symptoms in a Time of Fear Arch Intern Med, September 9, 2002; 162(16): 1809 - 1813. [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Zautra, A. J. Articles by Smith, B. W. Search for Related Content PubMed PubMed Citation Articles by Zautra, A. J. Articles by Smith, B. W. Related Collections Musculoskeletal Depression Psychophysiology