Interpersonal Stress, Depression, and Disease Activity in ... · tion (Kronfol et al., 1983;...

Health Psychology1994, Vol. 13, No. 2,139-148

Copyright 1994 by the American Psychological Association, Inc., andthe Division of Health Psychology/0278-6133/94/$3.00

Interpersonal Stress, Depression, and Disease Activity in RheumatoidArthritis and Osteoarthritis Patients

Alex J. Zautra, Mary H. Burleson, Kathy S. Matt, Sanford Roth, and Lisa Burrows

The relationships among interpersonal stressors, depression, coping inefficacy, hormones (prolac-tin, cortisol, and estradiol), and disease activity were examined. The sample comprised 33 womenwith rheumatoid arthritis (RAs; age 37-78) and 37 women with osteoarthritis (OAs; age 47-91),who served as controls. In a regression analysis, interpersonal conflict events accounted for morethan twice as much variance in depression in RAs than in OAs. In the RA patients, theimmune-stimulating hormones prolactin and estradiol were significantly positively correlated withinterpersonal conflicts, depression, coping inefficacy, and clinician ratings of disease activity,suggesting that RAs are more reactive to interpersonal stressors than are OAs, both psychologi-cally and physiologically.

Key words: interpersonal stress, arthritis, depression, hormones

Both life stress and psychological adjustment are thought toplay important roles in illness resistance and disease course forthose who become ill (Borysenko & Borysenko, 1982; Hinkle,1974; Kiecolt-Glaser et al., 1987; Locke, 1982; Rahe, 1974).Rheumatoid arthritis (RA) patients—who must cope withfunctional loss and recurrent, painful flare-ups from a chronicdisease—may be among the most vulnerable to the disruptiveeffects of life stress (Depue & Monroe, 1986). Indeed, thereare few illnesses potentially more debilitating than RA (Ander-son, Bradley, Young, McDaniel, & Wise, 1985). In its activestages, the disease is characterized by unpredictable episodesof extreme pain and disability, against a background of chronicimpairment. Because of the strain placed on their capacity toadjust psychologically, both from the illness itself and alsofrom related psychosocial stresses, RA patients may be at greatrisk for collapse of their physical health. A vicious cycle ofillness, stress, and illness recurrence may become established(Solomon, 1981). This study examines the effect of everydaystress on the mental and physical health of arthritis patients.

There has been considerable conjecture on the role ofpsychosocial stress in provoking flare-ups and disease progres-sion in RA. Stress is the cause most often given by RA patientsfor their flare-ups (Affleck, Pfeiffer, Tennen, & Fifield, 1987).Nevertheless, few studies have supported a direct link betweenlife stress and illness course in human RA subjects (cf.Hendrie, Paraskevas, Baragar, & Adamson, 1971; Zautra etal., 1989). Barriers to the demonstration of such effects may be

Alex J. Zautra, Mary H. Burleson, Kathy S. Matt, and Lisa Burrows,Department of Psychology, Arizona State University; Sanford Roth,The Arthritis Center, Inc., Phoenix, Arizona.

This research was supported in part by grants from the ArizonaDisease Control Research Commission (82-1707) and the GraduateCollege Research Assistantship Award Program at Arizona StateUniversity.

Correspondence concerning this article should be addressed to AlexJ. Zautra, Department of Psychology, Arizona State University,Tempe, Arizona 85287-1104.

found in shortcomings of measurement and methods on bothsides of the stress and illness equation.

The complexity of assessing and distinguishing among themany components of the life stress process has been aformidable obstacle. In fact, reliance on inadequate measure-ment instruments may have led researchers to dismiss prema-turely the role of stress in physical illness in studies where nosignificant effects of stress on health were observed (e.g.,Schroeder & Costa, 1984). On the other side of the equation,few have attempted to identify specific neuroendocrine andimmune parameters likely to mediate the effects of stress ondisease activity in RA. These shortcomings have left majorgaps in our understanding of the processes responsible for theobserved effects of life stress on the course of RA.

Psychosocial Stress and Rheumatoid Arthritis

Recently, advances have been made in the measurement ofeveryday life events that may add precision to the detection ofchanges in life stress associated with illness. Unlike majorevents, which occur infrequently, these small stressors fluctu-ate widely over short periods of time, resulting in meaningfuldifferences between subjects across days and weeks (Tennen,Suls, & Affleck, 1991). Thus, they provide a potentially bettermethod for testing the effects of life stress in illnesses such asRA, where disease flare-ups can occur and recur in a singlemonth.

Several investigators have suggested a vulnerability model inwhich small chronic stressors would play an important role indisease progression. In this model, the occurrence of smaller,everyday events in sufficient quantities may place enoughadditional stress on a person already weakened from chronicillness to precipitate an illness episode (Depue & Monroe,1986; Kessler, Price, & Wortman, 1985; Zautra, Guarnaccia, &Dohrenwend, 1986). Initial research has found support for thisvulnerability model by showing that everyday stressors areassociated with illness susceptibility (Delongis, Folkman, &Lazarus, 1988; Stone, Cox, Valdimarsdottir, Jandorf, & Neale,

139

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

140 ZAUTRA, BURLESON, MATT, ROTH, BURROWS

1987). However, research has yet to address the strength of theassociation between everyday stress and illness, to identifywhich stressors are the most salient, or to determine whichpatient groups are the most vulnerable.

There are many types of everyday stressors, and some formsof stress are more likely to be associated with distress anddisruption of internal homeostasis than others. In a diary studyof married couples, Bolger, Delongis, Kessler, and Schilling(1989) found that daily interpersonal stressors were, by far, themost distressing events. Among the chronically ill, ongoingstrains from a troubled marriage and other interpersonaltensions may be especially likely to cause increased psychologi-cal and biological disturbance. For example, family conflict hasbeen associated with poor adjustment to chronic illness (Coyne,Wortman, & Lehman, 1988), as well as to immune systemdysfunction (e.g., Kiecolt-Glaser et ah, 1987).

There is reason to suspect that RA patients may beespecially sensitive to interpersonal conflict events. They are atrisk for severe disability and are therefore dependent on themaintenance of ties with caregivers, even when those relation-ships are troubled (Manne & Zautra, 1989). In fact, poorpsychological adjustment to RA has been associated withtroubled family relations (Manne & Zautra, 1989; Shochet etah, 1969).

In addition to the stress-to-illness hypothesis, depression isalso important to examine. Not only are mental health out-comes significant in their own right, but the patient's psycho-logical condition may also play a pivotal role in diseaseprogression. From a biobehavioral perspective, the level ofdepressive symptoms may provide one of the best measures ofmental health. Patients with chronic illness and in pain arelikely to report many more depressive symptoms (Haythorn-thwaite, Sieber, & Kerns, 1991). Furthermore, stressful inter-personal environments are known to increase the frequency ofdepressive symptoms in both healthy and ill groups (Finch &Zautra, 1992). Depression has also been associated withendocrine changes (Heninger, Charney, & Sternberg, 1984;Miyabo, Asato, & Mizushima, 1977; Miyabo, Hisado, Asato,Mizushima, & Ueno, 1976) and lymphocyte regulatory dysfunc-tion (Kronfol et al., 1983; Schleifer, Keller, Siris, Davis, &Stein, 1985; Weisse, 1992). Indeed, the demoralization andfeelings of helplessness that accompany depression may be acritical risk factor in further exacerbations of RA.

Negative self-evaluation of coping efficacy is another factorrelated to both depression and life stress (Aldwin & Revenson,1988; Long, Chastain, Ung, Shoor, & Holman, 1989; Schiaffino& Revenson, 1992; Zautra & Wrabetz, 1991). Depressedpatients are likely to undervalue their capacity to cope effec-tively with disease and other stressful events. By doing so, theymay increase their risk of further illness exacerbations, thuscontributing to the vicious cycle of stress and disease progres-sion. Consistent with this argument, perceived inability to copewith future stressors has been associated with poorer mentalhealth and higher circulating B-cell proportions in RA patientsin our past studies (e.g., Zautra et al., 1989).

Physiological Mediators

To provide a comprehensive model of the effects of psycho-social stressors on health, it is important to understand how

these external signals exert their influence by changing internalmessages such as hormonal signals, which then alter theinternal physiological state. Much past research on stress hasfocused on the hypothalamic-pituitary-adrenal axis and itsrole in immune function (Selye, 1956). Stress, including socialstress (Meyerhoff, Oleshansky, & Mougey, 1988), affects thecentral nervous system, causing the eventual release of prolac-tin (PRL) and adrenal corticotrophic hormone (ACTH) fromthe pituitary gland and cortisol from the adrenal cortex(Euker, Meites, & Riegle, 1975; Matt, Scares, Talamantes, &Bartke, 1983; Neill, 1970; Plotsky, 1991). In the healthy indi-vidual, cortisol suppresses cellular immune function (thusreducing inflammation; Cohen & Crnic, 1982), whereas PRL isstimulatory (Russell, 1989). These opposing actions, as well asthe action of gonadal steroids, allow for modulation of theimmune response around a homeostatic set point.

However, because RA is a systemic autoimmune disorder,the homeostatic mechanisms that normally regulate immunityand inflammation may not be fully functional (Dorian &Garfinkel, 1987). Dysregulation in this system suggests thatRAs might be especially hormonally responsive to psychosocialstress, possibly leading to increased immune system activityand disease flare-ups. Immune activation in RA may also beaugmented through other immune-stimulating hormones, suchas estrogen. Estrogen is associated with increased autoanti-body production (Ahmed, Dauphinee, Montoya, & Talal,1989) and also stimulates PRL release (Wiedmann, Schwartz,& Frantz, 1976). Risk of RA is three times greater in womenthan in men, and autoimmune disease activity is known toincrease in the presence of estradiol (McGregor, 1990)—features consistent with the role of estrogen in the autoim-mune response. In summary, even though many factors besidesstress regulate adrenal and gonadal steroid hormone release(Spector, 1989), we thought it would be useful to examinecirculating levels of PRL, cortisol, and estradiol (estrogen),because these hormones could mediate the effects of stress anddepression on the immune system and disease activity in RA.

We designed the study to explore the effects of interpersonalstresses on depression and disease state in RA patients intreatment for their disease. Blood samples were collected andanalyzed for serum levels of PRL, cortisol, and estradiol.Patients with osteoarthritis (OA) in treatment were used ascontrols; these patients were expected to experience a moder-ate degree of pain and discomfort, like the RAs, because theywere seeking treatment. However, OAs differ from RAs in thatthey have a localized joint pain from cartilage damage, notfrom an immune-related inflammation such as occurs with RA.The diseases differ in clinical course as well as physiopathol-ogy. RA can be crippling and is progressive in its effectsleading to erosion in the bone itself, whereas osteoarthritis hasa much more benign clinical course (Lorig & Fries, 1990).There was good reason to believe, therefore, that the RAs andOAs would be similar in immediate concerns due to pain anddiscomfort but that the RAs would be more reactive toimmune system changes and perhaps more psychologicallysensitive to their social networks, because they could not beassured of any stability in their level of functional indepen-dence in the future.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

INTERPERSONAL STRESS AND ARTHRITIS 141

Method

Participants

Participants were 33 female RA patients and 37 female OA patients,who routinely came to an arthritis clinic for monthly blood tests tomonitor their medication. Ninety percent of the participants wereWhite, 3% were African American, and 7% were Hispanic. Approxi-mately 73% of the RA patients were married, 6% were single, 6% hadbeen widowed, and 15% were divorced. Among the OA patients, 60%were married, 5% were single, 24% had been widowed, and 11% weredivorced. Eighty-eight percent of the RAs and 86% of the OAs hadcompleted high school or beyond.

The RA patients were recruited from a list, prepared by the clinicstaff, of all female RA patients treated at the clinic. All were diagnosedas having either classical or definite RA (Rodnan & Schumacher,1983). The OA patients were recruited from among the femalepatients diagnosed with OA visiting the clinic on the same days thatthe RA participants were interviewed. Approximately one third of thepatients who were asked to participate in the study declined. Approxi-mately equal numbers of RAs and OAs refused to participate in thestudy.

Procedure

Patients were approached in the waiting room of the arthritis clinicat the time of their regularly scheduled appointments and asked if theywished to participate in the study. If the patient agreed to participate,she signed an informed-consent statement and was paid $10. She wasthen taken to a private room, in which a structured interview wasconducted. The interview included questions concerning demographicinformation, length of diagnosis with the disease, joint tenderness andswelling, amount of arthritis pain the patient was currently experienc-ing, and activity limitation. Additionally, patients were asked to reportrecent desirable and undesirable non-arthritis-related life and healthevents, interpersonal stressors, and positive interpersonal events.Coping efficacy for arthritis flare-ups and recent interpersonal stress-ors was also assessed. Patients then filled out a packet that containedmeasures of depression, psychological distress, and mood.

The structured interviews were suspended, if necessary, in order fora physician or a nurse practitioner at the arthritis clinic to examine thepatient. Physicians' joint assessments and global ratings of arthritisactivity were obtained at this time. The interviews were resumed oncompletion of the clinician's examination. If the patient was unable tostay at the clinic for the full interview, it was completed over thetelephone as soon as possible. Fifteen percent of the interviews werefinished over the telephone on the same day as the patient's visit to theclinic, and 5% were completed the following day.

Information on medications that the patient was currently takingwas obtained from the patient's medical records. Nursing staff at thearthritis clinic routinely verified medication use with each patient atthe time of their visit to the clinic. Thus, this information was accurateand up-to-date.

Nursing staff at the arthritis clinic were alerted to draw 10 ml beyondwhat was needed for the clinic's analysis. The blood was delivered tothe project biochemist for analysis.

Measures

Events. In this study, everyday desirable and undesirable eventswere measured with the Inventory of Small Life Events (Zautra et al.,1986). This inventory was constructed to minimize confounding withpersonality dispositions or affective states while preserving establishedcriteria for what constitutes a life event. All items on the list werewritten to have a discrete beginning, to be scorable as desirable or

undesirable (by a consensus of 80% or higher), and to refer toobservable changes in everyday activities rather than internal states.

Participants reported whether each event happened in the pastweek or past month and whether its occurrence was related to theirarthritic condition. They rated 144 small events (e.g., "got a trafficticket," "had a party"). These events were further subdivided intodesirable, undesirable, and ambiguous-desirability event categories.The events covered 13 different areas of life—including maritalrelations, children, extended family, social relations, recreation, school,work, finances, and health events.

It was expected that interpersonal events would be the most salientfor the patients. Therefore, the analyses focused on these events.Small events with interpersonal content were summed across catego-ries of marital relations, children, extended family, social relations,and recreation. Two scores were constructed from these summations:one for interpersonal conflict events and one for positive interpersonalevents. Noninterpersonal stressors were also examined for theircorrelations with depression, to check on our assumption that theinterpersonal events were the most salient for this population.

To reduce a possible confounding with illness-related stressors,events that the subject reported could be attributed, at least in part, totheir arthritic condition were not scored. Non-arthritis-related healthsymptoms were summed separately and were used in the analyses.

Coping inefficacy. Efficacy in coping with both nonarthritis eventsand arthritis flare-ups was also assessed, following the method used byLennon, Dohrenwend, Zautra, and Marbach (1990), Zautra et al.(1989), and Zautra and Wrabetz (1991). The patients were asked toselect two stressful events from those that had occurred during the pastmonth: the most stressful non-arthritis-related event and their mostrecent arthritis flare-up. They then answered 10 questions concerningtheir responses to the event, such as whether they attempted to "relaxand forget about what happened," or to "do something to prevent afuture recurrence." Immediately following the inquiry about specificcoping responses, the interviewer asked the patient to answer twoquestions using 7-point scales: (a) "How satisfied are you with how youresponded to the event," and (b) "If an event like this one were tohappen in the future, how certain are you that you would be able toadjust well to its negative aspects?" Coping inefficacy was computed asthe mean of scores on these two items separately for arthritis andnonarthritis stressors. Thus, coping inefficacy ratings were grounded inspecific coping responses made to the events and expectations aboutsimilar future events. The alpha reliability coefficients of copinginefficacy for the non-arthritis-related stressful event and the flare-upwere .48 and .64, respectively. Three subjects (2 RAs) reported nonegative events independent of arthritis and, therefore, did not havescores on coping with non-arthritis-related events.

Depression. The 10-item Depression subscale from the MentalHealth Inventory (Veil & Ware, 1983) was used to measure thisvariable. Participants used a 6-point scale to respond to items such asthe following: "How much of the time have you felt lonely," "Howoften did you feel nothing turned out the way you wanted it to," and"How much of the time have you been in low or very low spirits?" Themean across 10 items constituted the score on depression used in theanalyses. Past research has shown this measure to be internallyconsistent and stable over time and to accurately assess individualdifferences in psychological distress among RA patients and olderadults in general (Manne & Zautra, 1989; Zautra, Reich & Guarnac-cia, 1988). In this study, an alpha reliability coefficient of .91 wasobtained.

Self-reports of disability and pain. The Health Assessment Question-naire (HAQ) from Fries, Spitz, Kraines, and Holman (1980) was usedto measure the degree to which participants experienced arthritis-related activity limitation during the 7-day period before their visit tothe arthritis center. This is a 20-item, self-report instrument in whichthe subject uses a 4-point scale to rate the amount of difficulty sheexperiences when performing various daily tasks. The subject is asked

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.


Table 1Sample Characteristics: Illness Severity

Rheumatoid arthritis

Characteristic

Years with diagnosisActivity limitation (HAQ)Weighted joint countGlobal illness ratingCurrent pain

n

3333333333

M

13.981.888.452.24

37.45

SD

10.760.597.680.94

30.80

Osteoarthritis

n

3637373737

M

9.941.603.732.19

37.51

SD

8.520.513.720.81

24.56

t

-1.73-2.08*

3.21**-0.25

0.01

Note. HAQ = Health Assessment Questionnaire.*p < .05. **p < .01.

to assess her ability to perform daily tasks related to dressing andgrooming, arising, eating, walking, hygiene, household chores, andreaching and gripping household objects. The mean score for the 20items was computed. An alpha reliability coefficient of .93 wasobtained for this measure.

Patients were also asked to rate the amount of pain they werecurrently experiencing by means of a visual analogue scale rangingfrom no pain (0) topain as bad as it could be (100). The number of yearssince the patient had been diagnosed with her disease was also askedof the patient and corroborated with medical records.

Clinical ratings of disease activity. Physicians and nurse practition-ers used an abbreviated version of the American Rheumatism Associa-tion Medical Information System (ARAMIS) Joint Count (Fries &McShane, 1984) to assess the number of joints currently affected bydisease activity and the amount of disease activity in each joint.Previous studies have shown abbreviated versions of this measure to bevalid and reliable (Fuchs, Brooks, Callahan, & Pincus, 1989). Thefollowing 28 joints on the right and left sides of the body were includedin the assessment: distal interphalangeal (fingers), proximal interpha-langeal (fingers), metacarpophalangeal (first knuckle), carpometacar-pals, wrists, elbows, shoulders, sternoclaviculars, temporomandibulars(jaws), hips, knees, ankles, metatarsophalangeals (first foot knuckles),and proximal interphalangeal (toes). Current disease activity in eachof these joints was rated on a scale ranging from 0 to 3. Ratings ofdisease activity for each joint were summed to obtain a count of thenumber of affected joints weighted by disease activity (weighted jointcount). Clinicians also made a global assessment of current diseaseactivity by means of a 4-point scale ranging from no disease activity (0)to severe disease activity (3).

Hormone measurement. Blood was drawn from patients as previ-ously described, and the serum was stored at -50 °C until the time ofassay. Levels of PRL, cortisol, and estradiol were measured byradioimmunoassay with kits purchased from ICN Biomedicals, Inc.Samples were all measured in duplicate. Serum PRL was measuredwith a double antibody precipitation assay, and values are expressed inng/ml. The intra-assay coefficient of variation (CV) was 6.9%. Serumcortisol was measured directly in serum, using the coated tube assayprocedure of ICN Biomedical, Inc., and values are expressed as n-g/dl;the intra-assay CV was 9.6%. Serum estradiol was measured directly inthe serum, using a double antibody precipitation assay, and all valuesare expressed as ng/ml. The intra-assay CV was 11. 3%. All of thesample values fell within the following normal ranges: cortisol (morn-ing), 7-24 (ig/dl; cortisol (afternoon), 3-11 u.g/dl; PRL, 8-24 ng/ml;estradiol (postmenopausal), less than 0.03 ng/ml.

Because of problems in drawing and transporting blood samples tothe laboratory where analyses were performed, full hormone datawere available for only 22 of the 33 RA subjects and 28 of the 37 OAs.Therefore, the analyses reported for the psychosocial data alone havea larger sample size than the analyses that included endocrine data.However, the psychological results were also confirmed on the smallersample of subjects, for whom there was full data on hormone levels.

Results

Preliminary analyses were conducted for group differencesin demographic variables and illness severity. The OA and RAgroups did not differ in marital status or education. However,the OA subjects were older than the RAs (M = 65.86, SD =8.20, vs. M = 60.53, SD = 10.95), t(32) = 2.29, p < .05. Theyalso had slightly lower annual income (M = $35,000 vs.$27,500), t(33) = -2.57, p < .05, but subsequent analysesshowed that these income differences had nonsignificanteffects on the measures under study. The illness characteristicsof the participants are provided in Table 1. The OA patientsshowed less activity limitation and a lower weighted joint countthan did the RA patients. The clinician's global assessment ofdisease status did not differ between groups, and there were nodifferences between groups in level of present pain reportedon a visual analogue scale. RAs tended to report slightly moreyears with diagnosis than OAs, but the difference was notstatistically significant.

Table 2 contains analyses of differences between groups onthe major variables of interest in this study: life events, mentalhealth measures, and endocrine parameters. Endocrine valueswere log transformed to reduce the skewness of their distribu-tions. RAs differed from OAs in only one area: RAs reported0.4 fewer symptoms of ill health unrelated to their arthritisthan did OAs. The similarity on these measures indicates thatthe two groups are highly compatible in terms of life stress andpsychological distress and that they demonstrate no cleardisparity in mean levels of PRL, cortisol, and estradiol.

Patients were asked to identify all prescription and nonpre-scription medications that they were taking from a 62-itemchecklist. This list was corroborated by medical chart review.From these data, we determined that 6 RA patients weretaking low dosages of methotrexate (a disease-modifyingdrug), 14 were taking nonsteroidal anti-inflammatory drugs(NSAIDs) only, and 6 were taking a combination of NSAIDtherapy and methotrexate. Thirty-four of the OA patients weretaking NSAIDs. Additionally, 1 RA patient was on gold, 5were on low-dosage steroidal compounds, 2 were on antidepres-sive medications, and 11 were taking over-the-counter anti-inflammatory medications such as aspirin. Nine RAs and 14OAs were on estrogen-replacement therapy.

Before assessing the relationships between stress and dis-tress, we examined whether use of certain medications wasassociated with scores on interpersonal stress, coping ineffi-cacy, depression, and the endocrine measures: cortisol, PRL,

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.


Table 2Sample Characteristics: Life Events, Mental Health, and Endocrine Parameters

Rheumatoid arthritis Osteoarthritis

Characteristic M SD M SD

Event measuresTotal small undesirable eventsTotal positive eventsNonarthritis health symptomsInterpersonal conflict eventsPositive interpersonal events

Mental health measuresDepressionCoping inefficacy for non-arthritis-related

eventCoping inefficacy for flare-up

Endocrine parametersCortisolProlactinEstradiol

3333333333

33

3033

222222

4.24218.8180.6671.879

11.333

2.34

2.222.08

0.990.46

-1.16

3.2129.2310.7361.8675.560

0.78

0.940.89

0.300.550.41

3737373737

37

3435

282828

3.91921.1621.0811.541

12.622

2.30

2.282.06

1.130.57

-1.36

3.8188.4970.9542.0635.320

0.68

0.940.86

0.200.580.46

-0.381.112.02'

-0.720.99

-0.21

0.27-0.09

1.850.73

-1.61

"p < .05.

and estradiol. Medications were grouped into seven types foranalysis: antidepressive medications, low-dose steroids, nonste-roidal anti-inflammatory medications, disease-modifying medi-cations, ulcer medications, pain medications, and estrogen.Pearson product-moment correlations were computed, andone- or two-tailed tests (p < .10) were used, depending onwhether the direction of the relationship could be presumedon the basis of pharmacological effects of the drugs. Patientson estrogen replacement had higher scores on estradiol,r(60) = .39,;? < .01, but showed no significant differences onother measures in the study at/? < .10. Patients on low-dosesteroid medications had lower scores on cortisol, as expected,r(60) = — .36, p < .01, but that medication was uncorrelatedwith any other measure. Patients on antidepressive medica-tions reported more depression, r(60) = .27, p < .05; thatrelationship was most likely a consequence of the patient'sdepression leading the doctor to prescribe, rather than due tosome paradoxical effects of the drug. Partial correlationalanalyses and partial regressions were performed to test whethercontrolling for the linear effects of low-dose steroids orestrogen replacement would affect the relationships among the

variables; no significant effects were noted. The results arereported below, therefore, without controlling for these vari-ables.

Effects of Life Events and Coping Efficacy

The correlations between depression and scores on interper-sonal events, health problem events (not related to arthritis),and reports of coping inefficacy for RAs and OAs are shown inTable 3. As predicted, interpersonal conflict events wereassociated with higher scores on depression for both groups,and the effect was greatest in the RA sample. Noninterper-sonal negative events were unrelated to depression for bothgroups: r(37) = .18, p > .10, for OAs; r(33) = .11,p > .10, forRAs. Positive interpersonal events and inefficacy in copingwith stressful events other than arthritis had significant associa-tions with depression in both groups. Health stressors notrelated to arthritis were correlated with depression for the RAsample, but not for OAs.

To examine further the apparent difference between RAand OA groups in the size of the correlations, multiple

Table 3Correlations Between Depression and Interpersonal Events, Health Symptoms,and Coping Inefficacy

Depression in RAs Depression in OAs

Variable

Interpersonal eventsPositive3

Conflictualb

Nonarthritis health symptomsCoping inefficacy

Flare-upsOther events

333333

3232

-.49**.68***.31*

.28

.33*

373737

3636

-.48**.36*.07

.23

.32*

Note. RAs = subjects with rheumatoid arthritis, OAs = subjects with osteoarthritis."Partial correlations, controlling for interpersonal conflict events. bPartial correlations, controlling forpositive interpersonal events.*p < .05. **/> < .01. ***/? < .001.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.


Table 4Multiple Regression Prediction of Depression: Full Model

Step variables

CovariatesNonarthritis health symptomsActivity limitation (HAQ)Age

Main effectsInterpersonal conflict eventsPositive interpersonal eventsGroup (OA versus RA)Coping inefficacy for flare-ups

InteractionConflict Events x Group

Note. N = 68. HAQ = Health

R2 change

.21**

.19**

.08"

Assessment Questionnaire,

P

.119

.327

.004

-.290-.038-.093

.203

.439

OA =

SE

.101

.156

.008

.226

.016

.159

.080

.146

osteoarthritis, RA =

t

1.1822.092*0.414

-1.280-2.336*-0.584

2.546*

3.004**

rheumatoidarthritis.*p < .05. **p < .01.

regression analyses were performed with depression scores asthe criterion. Predictors included centered scores on interper-sonal events, scores on coping inefficacy during flare-ups, adummy variable to classify subjects according to illness group(RA vs. OA), and product terms representing the interactionsbetween interpersonal stress events and illness group andpositive interpersonal events and illness group. Coping ineffi-cacy with flare-ups was used in the equation, instead of copinginefficacy with other events, because it constituted a usefuladditional control over differences in depression resultingfrom differences in coping with illness-related stresses thatwere confounded with interpersonal stressors. In addition,group differences in age, activity limitation, and non-arthritis-related health events were controlled statistically by forcingthem into the regression equation first, before assessing theeffects of interpersonal events and their interactions withgroup status. The results are shown in Table 4.

There was a significant interaction effect in the prediction ofdepression. RAs showed higher elevations of depression thanOAs when experiencing interpersonal conflict events. Positiveinterpersonal events and coping efficacy for arthritis-relatedevents were associated with less depression in both OAs andRAs. There were no significant interactions between positiveevents or coping efficacy and illness group in the prediction ofdepression.

Effects of Stress and Depression on Endocrine Levels

If stress and poor coping efficacy are linked to diseaseactivity, their effects might be mediated by increases incirculating levels of hormones that amplify the immune re-sponse, such as PRL and estradiol. Log-transformed scores onPRL and estradiol were summed to create a composite indexof serum levels of immunostimulatory hormones. These scoreswere then correlated with measures of interpersonal conflictevents, positive interpersonal events, depression, and copinginefficacy. In addition, multiple correlation analyses wereperformed, to estimate the overall effect of psychosocialvariables on stimulatory hormones for RAs and OAs, sepa-rately. The results of these analyses are shown in Table 5.

As would be expected from a model in which the effects ofstress are mediated through changes in endocrine levels,higher interpersonal conflict was associated with higher levelsof immunostimulatory hormones for RAs, and the positiverelationship between depression and these hormones ap-peared to be even greater. Coping inefficacy was positivelyrelated to levels of immune-stimulating hormones and also todepression. In summary, those RA patients with higher scoreson depression, and those who felt they coped poorly withstressful events, showed higher levels of stimulatory hormones.The findings were comparable when analyzing correlates of

Table 5Correlations Between Stimulatory Hormones (Prolactin + Estradiol) and Psychosocial Factorsand Clinician Ratings of Disease

Variable Rheumatoid arthritis Osteoarthritis

Psychosocial factorsInterpersonal conflict eventsPositive interpersonal eventsDepressionCoping inefficacy with flare-upsCoping inefficacy with other events

Multiple RClinician ratings of disease

Weighted joint countGlobal illness assessment

.384*

.227

.493*

.543**

.520**

.763*

.423*

.433*

.136

.143

.085

.080

.217

.260

.238

.311

For subjects with rheumatoid arthritis, n = 22. For subjects with osteoarthritis, n = 28.•p < .05. *"/> < .01.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.


Osteoarthritis Patients, N=28X 2 (3) = 6.775, p = .079TLI a -.592

Interpersonalconflictevents

.193

*

*)<

.996 *.981 ^ t

f f

Depression.085 Stimulatory

hormones.311 Global

IllnessAssessment

Rheumatoid Arthritis Patients, N=22X2(3) = 4.186, B=-242 .901TLI = .842

Interpersonalconflictevents

.853.s

.521*If-

Depression.493* ^

Stimulatoryhormones

.433**

GlobalIllnessAssessment

* E < .01

Figure 1. Path models for osteoarthritis patients and rheumatoid arthritis patients. Parameter estimatesare standardized. Disturbances are represented by diagonal arrows. TLI = Tucker-Lewis index forgoodness of fit.

PRL and estradiol separately, with the exception of positiveinterpersonal events, which were significantly correlated withPRL,r(22) = .42,p < .05, but not estradiol, r(22) = -.15,p >.10.

Similar effects were not observed for OA patients; neitherdepression nor coping efficacy was related to stimulatoryhormone levels for PRL and estradiol. Tests of differences inthe size of correlations were nonsignificant. Interaction effectsbetween any single variable and group status (RA vs. OA) alsoproved to be nonsignificant, although differences betweenOAs and RAs approached significance for coping inefficacywith flares in the expected direction, t(46) = l.76,p < .09. Ananalysis of differences in overall effects due to the compositesof psychosocial influences was conducted by comparing mul-tiple Rs obtained from the multiple correlational results(Tabachnick & Fidell, 1989). The overall difference in multipleRs was significant, z-difference score = 2.405,p < .025.

Correlational analyses were also performed for log trans-formed scores on cortisol. These findings were more ambigu-ous. For OA subjects, cortisol scores were not correlated withany of the psychosocial measures. For RA patients, neithercoping inefficacy nor stressful events were significantly corre-lated with elevations in cortisol. Depression, however, wasrelated to cortisol, r(22) = .37,p < .05.

Hormones Levels and Disease Status

To complete the test of our mediational model, hormonelevels were correlated with assessments of disease status. AsTable 5 shows, both physician's joint count and physician'sglobal assessment of disease state were associated positivelywith levels of immune-stimulating hormones. These effectswere significant in the RA sample and approached significancein the OAs. Cortisol levels in OA patients were uncorrelated

with clinical ratings of disease activity (ps > .10). In RApatients, however, cortisol was correlated with higher (and notlower) clinician ratings of disease activity, r(22) = AS,p < .05,and was marginally correlated with joint count, r(22) = .33,p < .15.

Path Model

RA patients with greater interpersonal conflict showed apropensity toward greater depression, which in turn wasassociated with elevations in PRL, estradiol, and cortisol.These hormone elevations were in turn associated with evi-dence of greater disease activity. To examine this mediationalmodel more formally, we conducted a path analysis, using thecovariance matrices of relationships among scores on fourvariables for each group: interpersonal conflict events, depres-sion, immune-stimulating hormones, and the clinician's globalrating of disease activity. We limited the number of variablesinput to this analysis because of the small sample sizesinvolved. Coping inefficacy was left out of this analysis, eventhough it is an important variable, because that variableoverlapped with depression in their relationship to hormonelevels. A fully mediated effects model was tested for fit to thedata for each group, with the size of the parameters relatingthe constructs free to vary among the groups. The results areshown in Figure 1.

The model fit the data well for the RAs, as evidenced by thenonsignificant chi-square statistic, a Tucker-Lewis goodness-of-fit index of .842, and significant path coefficients. Themediated model fit less well for the OA group, with amarginally significant chi-square and a poor Tucker-Lewisindex. The small size of each sample precluded a directstatistical test of differences between models. Nevertheless, animportant difference between the two groups was that all of

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.


the parameters were statistically significant for the RA group,supporting a fully mediated model for the effects on interper-sonal conflict events, but none of the parameters were signifi-cant for the OAs.

Discussion

The purpose of this investigation was to explore differencesbetween RA and OA patients while testing the influences ofpsychosocial stressors on depression, hormone levels, anddisease manifestations. The findings suggest that these twopopulations of arthritis patients do not differ in levels of stressand distress associated with their illness but may differ in theimportance of stress and coping to psychological adjustmentand also to physical health.

The two groups were quite similar on levels of pain, lifestress events, positive events, and global assessments of theseverity of their arthritis. They also showed no differences inlevel of depression or in their assessments of their efficacy incoping with arthritis and other stressful events. Furthermore,there were no significant differences in levels of PRL, cortisol,or estradiol between OAs and RAs. The overall similarity maybe attributable to sampling frame, because all subjects wererecruited from the waiting room of a large rheumatologypractice and therefore all were likely to be symptomatic. Forthis study, the lack of group differences on these majorvariables was advantageous. It allowed us to compare the twogroups' psychological and endocrine responses to psychosocialstress.

The differences between groups that were obtained wereexpected. OAs were older, showed less overall disability, andhad fewer joints involved in their disease than did RAs. Thesefindings are consistent with medical facts about the illnesses:RA onset is earlier in life and is systemic, often attacking anumber of joints bilaterally, in contrast to OA. The greaterdisability in RAs found in this study may be directly attribut-able to multiple joint involvement and the progressive natureofRA.

The relationship between interpersonal events and depres-sion was substantial in both groups, but the proportion ofvariance accounted for was particularly dramatic in the RAs.Thus, although the mean levels of interpersonal conflict anddepression did not differ in RAs and OAs, the relationshipbetween them did. In keeping with prior studies by Jamisonand Virts (1990), Manne and Zautra (1989), and Revensonand Felton (1984), those with stressful social interactions inthe past month displayed greater depression, and indepen-dently, those with positive social events were less likely to showdepressive symptoms. Individual differences in the frequencyof other stressful events appeared to have little or no effect ondepression, in comparison with the effects observed for inter-personal events. These results underscore the importance ofthe interpersonal arena among RA patients. They furthersuggest that future studies may benefit from the inclusion of acomprehensive inventory of interpersonal events such as wasused here and possibly excluding the bulk of events dealingwith noninterpersonal changes, to reduce the measurementburden on participants.

Why should RAs display greater reactivity to stressful

interpersonal events than their counterparts in the OA popula-tion? We suspect that degree of attachment to the supportsystem and expectancy of future dependence, both unmea-sured in this study, may contribute to the differences obtainedfor negative interpersonal events. Because RA is a progressivedisorder, the RA patient expects a decline in independentfunctioning, and this may heighten concern for a stable andsupportive relationship, beyond that likely for the OA group.Parenthetically, it is not surprising that positive events did notvary in their effects between groups. Heightened sensitivity, ifit is present, would most likely occur in response to thoseevents that threaten to disrupt the social equilibrium, and notevents that would enrich and enhance feelings of well-being.

Although the data are correlative, they also support the ideathat RAs are more physiologically sensitive to psychosocialstress than are OA patients. RA patients showed positiverelationships between the immunostimulatory hormones anddepression, interpersonal conflict events, and coping ineffi-cacy, whereas OA patients did not. These larger responses inRAs may be due to disease-related endocrine disregulation, orthey may be related to the increased psychological responsivityalready noted. Further study is needed to examine thesepossibilities. We hasten to add, however, that a distinctionbetween psychological vulnerability and physiological sensitiv-ity may be artificial; both body and mind may be more reactiveto stress and emotional upset in the RA patient in comparisonwith controls.

The path analyses support a mediational model of theeffects of interpersonal stress on the health of the RA patient.Note that the good fit of the mediational model does not meanthat another model would not fit the data equally well. Indeed,it is possible that a reverse causal ordering of the variableswould not lead to a large decrement in the goodness of fit ofthe data to the model. Nevertheless, the model fits ourunderstanding of the nature of the disease, and how patientsmay become affected by the psychosocial sequelae of thischronic illness. Over the course of RA, flare-ups of the diseaseinevitably lead to joint damage and greater disability, andhence to greater dependence on caregivers. Thus, the pressureto cope with interpersonal demands is likely to be particularlytroublesome for the RA patients, leading to increased depres-sion and a reduced sense of efficacy. These factors, contribut-ing possibly to an underlying cognitive schema of helplessness(Smith & Wallston, 1992), would then lead to a chronicallystressful internal state, which would elevate PRL and estradiollevels. These hormones are known for their stimulatory effecton immune cells. They may in turn activate key immunemechanisms associated with greater inflammatory response,leading to a disease flare.

The findings presented here are not definitive. Indeed, thisinitial study may be viewed as one that poses hypotheses ratherthan testing them. The cross-sectional nature of the studylimits the confidence we can have in the direction of therelationships implied in the path models and in the regressionanalyses. Furthermore, all possible effects of differences be-tween groups in age and medications could not be examined;complex interactions may take place between exogenous hor-mones and the physiology of the patient with autoimmunedisease that leads to changes in hormonal reactivity in ways not

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.


yet understood. As our subjects were patients in an ongoingclinical practice, it was neither feasible nor ethical to manipu-late their treatment, to estimate the total effects of medica-tions.

This study also did not examine those immune systemparameters that might forge a link between elevations inimmune-stimulating hormones and joint inflammation. In-deed, we cannot be certain that the hormones we expected tobe immune stimulating actually affected the immune param-eters in the way we assumed. Estradiol, for example, althoughgenerally regarded as stimulatory of autoimmunity (Nelson &Steinberg, 1987), has been regarded by some as potentiallyinhibitory of some immune functions (e.g., Homo-Delarche etal., 1991) and may have different effects on RA patients thananticipated. Likewise, cortisol, long regarded as immunosup-pressant, may not have its expected effect on patients with RA.The longitudinal assessment of change over time in theparameters under study here, with the addition of immunemarkers likely to covary with illness worsening such as interleu-kin-2 receptor scores (Harrington et al., 1993), would appearto be the next logical step in future research.

Even with its limitations, this study represents an importantadvance in research on this topic. For the first time, media-tional linkages were established from a variable outside thebody (stressful event measurement) through changes in inter-nal messengers (immune-stimulating hormone levels) to dis-ease state (clinician ratings of disease activity). Past studieshave been able to establish correlations between exogenousvariables and biological states such as hormonal levels orimmune parameters or have identified relations betweenbiological markers of disease activity and clinical levels ofdisease. To our knowledge, this is the first study of RA that hassimultaneously demonstrated significant relationships at bothlinks in this chain.

References

Affleck, G., Pfeiffer, C., Tennen, H., & Fifield, J. (1987). Attributionalprocessess in rheumatoid arthritis patients. Arthritis and Rheuma-tism, 30, 927-931.

Ahmed, S. A., Dauphinee, M. J., Montoya, A. I., & Talal, N. (1989).Estrogen induces normal murine CD5+ B cells to produce autoanti-bodies. Journal of Immunology, 142, 2647-2653.

Aldwin, C. M., & Revenson, T. A. (1988). Does coping help? Areexamination of the relation between coping and mental health.Journal of Personality and Social Psychology, 53, 337-348.

Anderson, K. O., Bradley, L. A., Young, L. D., McDaniel, L. K., &Wise, C. M. (1985). Rheumatoid arthritis: Review of psychologicalfactors related to etiology, effects, and treatment. PsychologicalBulletin, 98, 358-387.

Bolger, N., DeLongis, A., Kessler, R. C., & Schilling, E. A. (1989).Effects of daily stress on negative mood. Journal of Personality andSocial Psychology, 57, 808-818.

Borysenko, M., & Borysenko, J. (1982). Stress, behavior and immunity:Animal models and mediating mechanisms. General Hospital Psychia-try, 4, 59-67.

Cohen, J. J. & Crnic L. S. (1982). Glucocorticoids, stress, and theimmune response. In D. R. Webb (Ed.), Immunopharmacology andthe regulation of leukocyte function (pp. 61-91). New York: MarcelDekker.

Coyne, J. C., Wortman, C. B., & Lehman, D. R. (1988). The other side

of support: Emotional overinvolvement and miscarried helping. InB. Gotlieb, (Ed.), Social support: Formats, processes and effects (pp.305-330). Newbury Park, CA: Sage.

Delongis, A., Folkman, S., & Lazarus, R. S. (1988). The impact of dailystress on health and mood: Psychological and social resources asmediators. Journal of Personality and Social Psychology, 54, 486-495.

Depue, R. A., & Monroe, S. M. (1986). Conceptualization andmeasurement of human disorder in life stress research. PsychologicalBulletin, 99, 36-51.

Dorian, B., & Garfinkel, P. E. (1987). Stress, immunity & illness—Areview. Psychological Medicine, 17, 393—407.

Euker, J. S., Meites, J., & Riegle, G. D. (1975). Effects of acute stressof serum LH and PRL in intact, castrated, and dexamethasone-treated male rats. Endocrinology, 96, 85.

Finch, J. F., & Zautra, A. J. (1992). Testing latent longitudinal modelsof social ties and depression among the elderly: A comparison ofdistribution-free and maximum likelihood estimates with nonnor-mal data. Psychology and Aging, 7, 107-118.

Fries, J., & McShane, S. (1984). ARAMIS (the American RheumatismAssociation Medical Information System): A prototypical nationalchronic disease data bank. The Western Journal of Medicine, 145,798-804.

Fries, J., Spitz, P., Kraines, R., & Holman, H. (1980). Measurement ofpatient outcome in arthritis. Arthritis and Rheumatism, 23, 137-145.

Fuchs, H. A., Brooks, R. H., Callahan, L. F., & Pincus, T. (1989). Asimplified twenty-eight joint quantitative articular index in rheuma-toid arthritis. Arthritis and Rheumatism, 32, 531-537.

Harrington, L., Affleck, G., Urrows, S., Tennen, H., Higgins, P.,Zautra, A., & Hoffman, S. (1993). Temporal covariation of solubleinterleukin-2 receptor levels, daily stress, and disease activity inrheumatoid arthritis. Arthritis and Rheumatism, 36, 199-203.

Haythornthwaite, J. A., Sieber, W. J., & Kerns, R. D. (1991).Depression and chronic pain experience. Pain, 46, 177-184.

Hendrie, H. C., Paraskevas, F. D., Baragar, F. D., & Adamson, J. D.(1971). Stress, immunoglobulin levels and early polyarthritis. Jour-nal of Psychosomatic Research, 15, 337-342.

Heninger, G., Charney D., & Sternberg, D. (1984). Serotonergicfunction in depression. Archives of General Psychiatry, 41, 398-402.

Hinkle, L. E., Jr. (1974). The effects of exposure to cultural change,social change and changes in interpersonal relations on health. InB. S. Dohrenwend & B. P. Dohrenwend (Eds.), Stressful life events(pp. 9^t4). New York: Wiley

Homo-Delarche, F., Fitzpatrick, F., Christeff, N., Nunez, E. A., Bach,J. F., & Dardenne, M. (1991). Sex steroids, glucocorticoids, stressand autoimmunity. Journal of Steroid Biochemical Molecular Biology,40, 619-637.

Jamison, R. N., & Virts, K. L. (1990). The influence of family supporton chronic pain. Behaviour Research and Therapy, 4, 283-287.

Kessler, R. C, Price, R. H., & Wortman, C. B. (1985). Social factors inpsychopathology. American Review of Psychology, 36, 531-572.

Kiecolt-Glaser, J. K., Fisher, L. D., Ogrocki, P., Stout, J. C., Speicher,C. E., & Glaser, R. (1987). Marital quality, marital disruption, andimmune function. Psychosomatic Medicine, 49, 13-33.

Kronfol, Z., Silva, J., Jr., Greden, J., Dembinski, S., Gardner, R., &Carroll, B. (1983). Impaired lymphocyte function in depressiveillness. Life Sciences, 33, 241-247.

Lennon, M. C., Dohrenwend, B. P., Zautra, A. J., & Marbach, J. J.(1990). Coping and adaptation to facial pain in contrast to otherstressful life events. Journal of Personality and Social Psychology, 59,1040-1050.

Locke, S. E. (1982). Stress, adaptation and immunity: Studies inhumans. General Hospital Psychiatry, 4, 49-58.

Lorig, K., Chastain, R. L., Ung, E., Shoor, S., & Holman, H. R. (1989).Development and evaluation of a scale to measure perceived

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.


self-efficacy in people with arthritis. Arthritis and Rheumatism, 32,37-44.

Lorig, K., & Fries, J. F. (1990). The arthritis helpbook (3rd ed.).Reading, MA: Addison-Wesley.

Manne, S., & Zautra, A. (1989). Spouse criticism and support: Theirassociation with coping and adjustment to rheumatoid arthritis.Journal of Personality and Social Psychology, 56, 608-617.

Matt, K. S., Scares, M. J., Talamantes, F., & Bartke, A. (1983). Effectsof handling and ether anesthesia on serum prolactin levels in thegolden hamster. Proceedings of the Society of Experimental Biologyand Medicine, 173, 463-466.

McGregor, Alan M. (1990). Immunoendocrine interactions and auto-immunity. New England Journal of Medicine, 322, 1739-1741.

Meyerhoff, J. L., Oleshansky, M. A., & Mougey, E. H. (1988).Psychologic stress increases plasma levels of prolactin, cortisol, andPOMC-derived peptides in man. Psychosomatic Medicine, 50, 295-303.

Miyabo, S., Asato, T., & Mizushima, N. (1977). Prolactin and growthhormone responses to psychological stress in normal and neuroticsubjects. Journal of Clinical Endocrinology and Metabolism, 44,947-951.

Miyabo, S., Hisado, T., Asato, T., Mizushima, N., & Ueno, K. (1976).Growth hormone and cortisol responses to psychological stress:Comparison of normal and neurotic subjects. Journal of ClinicalEndocrinology and Metabolism, 42, 1158-1162.

Neill J. D. (1970). Effect of "stress" on serum prolactin and luteinizinghormone levels during the estrous cycle of the rat. Endocrinology, 87,1192-1197.

Nelson, J. L., & Steinberg, A. D. (1987). Sex steroids, autoimmuniry,and autoimmune diseases. In I. Berczi & K. Kovacs (Eds.), Hor-mones and immunity (pp. 92-119). Lancaster, England: MTP Press.

Plotsky, P. M. (1991). Neural coding of stimulus-induced ACTHsecretion. In M. R. Brown, G. F. Koob, & C. Rivier (Eds.), Stress,neurobiology, and neuroendocrinology (pp. 137-150). New York:Marcel Dekker.

Rahe, R. H. (1974). The pathway between subjects' recent life changesand their near future illness reports. In B. S. Dohrenwend & B. P.Dohrenwend (Eds.), Stressful life events: Their nature and effects (pp.73-86). New York: Wiley.

Revenson, T. A., & Felton, B. J. (1984). Disability and coping aspredictors of psychological adjustment to rheumatoid arthritis.Journal of Consulting and Clinical Psychology, 57, 344-348.

Rodnan, G. P., & Schumacher, H. R. (1983). Primer on the rheumaticdiseases (8th ed., pp. 207-208). Atlanta, GA: Arthritis Foundation.

Russell, D. H. (1989). New aspects of prolactin and immunity: Alymphocyte-derived prolactin-like product and nuclear protein ki-nase C activation. Trends in Pharmacologic Science, 10, 40-44.

Schiaffino, K. M., & Revenson, T. A. (1992). The role of perceivedself-efficacy, perceived control, and causal attributions in adaptationto rheumatoid arthritis: Distinguishing mediator from moderatoreffects. Personality and Social Psychology Bulletin, 18, 709-718.

Schleifer, S. J., Keller, S. E., Siris, S. G., Davis, K. L., & Stein, M.(1985). Depression and immunity: Lymphocyte function in ambula-tory depressed patients, hospitalized schizophrenic patients, andpatients hospitalized for herniorraphy. Archives of General Psychia-try, 42, 129-133.

Schroeder, D. H., & Costa, P. T. (1984). Influence of life event stresson physical illness. Journal of Personality and Social Psychology, 46,853-863.

Selye, H. (1956). The stress of life. New York: McGraw-Hill.Shochet, B. R., Lisansky, E. T., Schubart, A. F., Fiocco, V., Kurland,

S., & Pope, M. (1969). A medical-psychiatric study of patients withrheumatoid arthritis. Psychosomatics, 10, 271-279.

Smith, C. A., & Wallston, K. A. (1992). Adaptation in patients withchronic rheumatoid arthritis: Application of a general model. HealthPsychology, 11, 151-162.

Solomon, G. F. (1981). Emotional and personality factors in the onsetand course of autoimmune disease, particularly rheumatoid arthri-tis. In R. Ader (Ed.), Psychoneuroimmunology (pp. 159-184). SanDiego, CA: Academic Press.

Spector, T. D. (1989). Sex hormone measurements in rheumatoidarthritis. British Journal of Rheumatology, 2S(Suppl. 1), 62-68.

Stone, A. A., Cox, D. S., Valdimarsdottir, H., Jandorf, L., & Neale,J. M. (1987). Evidence that secretory IgA antibody is associated withdaily mood. Journal of Personality and Social Psychology, 52, 988-993.

Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (p.158). New York: Harper & Row.

Tennen, H., Suls, J., & Affleck, G. (1991). Personality and dailyexperience: The promise and the challenge. Journal of Personality,59, 313-338.

Veit, C. T., & Ware, J. E., Jr. (1983). The structure of psychologicaldistress and well-being in general populations. Journal of ConsultingClinical Psychology, 51, 730-742.

Weisse, C. S. (1992). Depression and immunocompetence: A review ofthe literature. Psychological Bulletin, 111, 475-489.

Wiedmann, E., Schwartz, E., & Frantz, A. (1976). Acute and chronicestrogen effects on somatomedic activity, growth hormone, andprolactin in man. Journal of Clinical Endocrinology and Metabolism,42, 943.

Zautra, A. J., Guarnaccia, C. A., & Dohrenwend, B. P. (1986).Measuring small life events.American Journal ofCommunity Psychol-ogy, 14, 629-655.

Zautra, A. J., Okun, M., Robinson, S., Lee, D. Roth, S., & Emmanuel,J. (1989). Life stress and lymphocyte alterations among rheumatoidarthritis patients. Health Psychology, 8, 1-14.

Zautra, A. J., Reich, J. W., & Guarnaccia, C. A. (1988). Some everydaylife consequences of disability and bereavement for older adults.Journal of Personality and Social Psychology, 59, 550—561.

Zautra, A. J., & Wrabetz, A. B. (1991). Coping success and itsrelationship to psychological distress for older adults. Journal ofPersonality and Social Psychology, 61, 801-810.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

Interpersonal Stress, Depression, and Disease Activity in ... · tion (Kronfol et al., 1983;...

Documents

Transcript of Interpersonal Stress, Depression, and Disease Activity in ... · tion (Kronfol et al., 1983;...