Relationship between appearance and psychological distress in rheumatic diseases




To examine the relationship between physical appearance concerns and psychological distress in patients with rheumatic diseases.


A total of 60 patients with systemic lupus erythematosus (SLE), 44 with chronic rheumatoid arthritis (RA), and 53 with recent-onset RA were evaluated for levels of appearance concern and a range of illness-specific measures to determine how these demographic and clinical variables were related to the dependent variable psychological distress.


Using hierarchical multiple regression analyses, we found that both appearance concerns and levels of disability were predictive of depression in patients with RA. In the SLE sample, physical disability was predictive of depression when appearance concerns were not included in the analysis. However, disability did not predict depression when appearance concerns were entered into the analysis. This indicates that appearance concerns mediated the relationship between disability and depression in SLE. There was no association between appearance concerns and anxiety in either sample.


The results suggest that appearance concerns are strongly related to depression in patients with rheumatic diseases and should be routinely assessed.


Rheumatic diseases, including rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE), are autoimmune diseases characterized by persistent inflammatory reactions, especially of the joints, and are accompanied by pain and fatigue. RA is an inflammatory disease that is systemic in nature but produces its most prominent manifestations in the joints, often leading to progressive destruction and deformity of cartilage and bone. Patients recently diagnosed with RA have fewer outward signs than do patients with chronic RA, who often have marked physical deformities, especially of their hands and feet. SLE is a chronic inflammatory autoimmune disorder that affects multiple organ systems. Although the most serious symptoms of SLE involve internal organ damage, SLE causes physical changes in appearance, most often rashes and lesions on the skin and hand deformity. Physical changes in SLE tend to be temporary during periods of disease activity.

Although RA and SLE are associated with perceptible changes to physical appearance, body image concerns have received little empirical investigation. However, research confirms high rates of disturbance to body image. It has been demonstrated that women with RA have lower levels of self-esteem compared with healthy women (1) and high levels of body image dissatisfaction (2). Indeed, research has found that women with SLE and RA have poorer body image than healthy women (3). The results of the study by Cornwell and Schmitt demonstrated that the body image disturbance for each woman corresponded to the part of her body most affected by her disease, suggesting that specific concerns are related to disfigurements (3). Nonetheless, disfigurement is not the sole determinant of body image dissatisfaction. Vamos and colleagues (4) found that all participants viewed their hands as fundamentally unattractive compared with their premorbid attractiveness. However, when women's subjective attractiveness ratings were compared with objective observer ratings, correlations were not significant (4). Therefore, physical disfigurement alone cannot explain the women's perceptions of their own attractiveness.

Psychosocial issues, including attitudes, psychopathology, and social functioning, might also be at least as strongly associated with body image disturbance as physical factors. Indeed, research involving women with scleroderma (another rheumatic disease associated with changes to the appearance of the skin) found that impaired psychosocial functioning and depression were associated with increased body image dissatisfaction and that dissatisfaction mediated the relationship between depression and psychosocial functioning in patients with scleroderma (5). These results supported a model in which dissatisfaction with appearance was associated with depressive symptoms, which were in turn related to poorer psychosocial functioning.

Confirming the relationship between depression, appearance concerns, and psychosocial functioning in other rheumatologic conditions would have important clinical implications. Research has found high rates of depression in patients with RA (6, 7) and SLE (8–10). Elevated distress has consequences for patients' function (disability) and quality of life. Research suggests that depression is more strongly associated with disability than disease parameters (11). By investigating which factors affect distress, psychological therapies can be improved to target specific areas and minimize the psychological impact of the disease.

The goal of the present study was to investigate the relationship between appearance concerns and psychological distress in 3 samples of patients with rheumatic disease. It was hypothesized that in patients with RA, appearance concerns would mediate the relationship between disability and psychological distress. For SLE, we used the physical domain of a health-related quality of life (HRQOL) measure. This domain measures the degree to which disability impacts upon physical role function and is strongly associated with disability. Therefore, we made the same predictions for patients with SLE, that appearance concerns would mediate the relationship between the physical HRQOL and distress.



A total of 157 patients, ages 18–75 years, from 3 clinical samples were included in the study. All participants could read and write in English and none had a history of psychiatric illness or substance abuse. For the SLE group, 60 patients who had been diagnosed with SLE by their treating consultant rheumatologist according to American College of Rheumatology criteria (12) were recruited. Consecutive patients attending a rheumatology outpatient clinic were asked to volunteer (n = 16). One author (Leslie Schrieber) confirmed the diagnosis for these patients. Volunteers were also recruited from an advertisement in the Lupus Association of New South Wales newsletter. In these cases, the treating rheumatologist was contacted to confirm disease status. Therefore, the SLE group was a convenience sample. A total of 62 individuals volunteered for the study. The diagnosis was not confirmed for 2 patients and they were excluded. Ninety-seven percent of the patients in this sample were women, with a mean ± SD age of 44.4 ± 12.15 years and a mean ± SD time since diagnosis of 10.19 ± 8.69 years.

The recent-onset RA sample of 53 patients was obtained from a consecutive sample of new referrals to rheumatology clinics in or near London (13). Participants were patients who had been diagnosed with definite or classic RA by a consultant rheumatologist within the past 2 years and who were seropositive for rheumatoid factor. Sixty-three patients were identified as meeting the inclusion criteria and 56 volunteered. Of these, 3 were later excluded due to changes in their diagnosis. Seventy percent of the participants were women, with a mean ± SD age of 55.06 ± 14.07 years and a mean ± SD duration of illness of 12.63 ± 8.22 months.

Fifty consecutive patients diagnosed with chronic RA by their consulting rheumatologist were approached in 3 hospital-based rheumatology clinics in and around Sydney, Australia. A total of 44 (88%) patients volunteered. Seventy-five percent were women, with a mean ± SD age of 58.63 ± 11.89 years and a mean ± SD time since diagnosis of 15.5 ± 12.8 years. Both patients with chronic RA and those with recently diagnosed RA completed the same assessments. Preliminary analyses determined that the pattern of relationships between the variables in these samples were similar. Therefore, for the purposes of analyses, the RA samples have been combined.


The participants in this study had volunteered for treatment programs as part of 3 randomized controlled trials and were assessed in a psychology or rheumatology clinic prior to intervention. Participants attended the clinic at a convenient time to complete the questionnaires, and patients with RA underwent joint assessments and had blood taken at the same appointment where possible or as soon as practicable afterwards (in all cases within 1 week). All participants gave informed consent to take part in the research, as approved by the relevant ethics committees.


Demographic variables were collected and participants completed the Disease Repercussion Profile (DRP) (14) and the Hospital Anxiety and Depression Scales (HADS) (15). The 2 samples completed a different set of questionnaires, which was necessary because many RA measures are not validated in or are irrelevant to an SLE sample. The RA participants completed the Stanford Health Assessment Questionnaire (HAQ) (16) and the Coping Strategy Questionnaire (CSQ) (17) and underwent joint assessments administered by a trained rheumatology nurse to measure swollen and painful joints. Assessors were blind to the questionnaire responses of participants. Additionally, the blood of patients with RA was analyzed for erythrocyte sedimentation rate (ESR) and C-reactive protein level (CRP). Patients with SLE were not given a physical test because a suitable measure is not available. Patients with SLE completed the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) to measure their HRQOL.

Patient and disease characteristics.

For both samples, the following demographic information was collected: sex, age, duration of illness (measured in years as time since diagnosis), living arrangements (indicating whether participants were married or cohabiting), level of education (in years of education), and employment (indicating whether the person worked full time, part time, or not at all). For the SLE sample, whether the person was currently in a flare of illness was also recorded (i.e., disease status).

Disease activity.

The Ritchie Articular Index (18), a widely used measure of joint function, is a detailed, sensitive measure of the number of actively inflamed and tender joints, and is administered by a trained rheumatology nurse. An indication of current disease activity for RA was obtained by analyzing blood samples for ESR and CRP.


The HAQ (16) was used to measure the levels of limitation and disability associated with illness. Excellent reliability and predictive validity have been reported for the HAQ (16).

Health-related quality of life.

The SF-36 is one of the most widely used questionnaires for assessing HRQOL in patients with SLE (16). The SF-36 has 2 measures: a physical composite score (PCS) and a mental composite score (MCS). The former measures the degree to which the illness impacts on the person's ability to complete major life roles, and the latter reflects how emotional functioning is affected by illness. While HRQOL is different from disability, the physical function subscale of the SF-36 is strongly related to disability (19).

Psychological distress.

Anxiety and depression were measured using the HADS (15). This questionnaire was specifically designed for use with physically ill patients and avoids reliance on the somatic symptoms of depression. A separate score is obtained for anxiety and depression.

Perceived appearance.

Appearance concerns were measured using question 6 of the DRP (14). Participants were asked, “How does your arthritis/lupus make you feel about yourself and your appearance at the moment?” Answer options were attractive, no effect, or unattractive. Participants who indicated a negative impact were then asked to estimate the importance of this impact on a visual analog scale (range 0–10).

Coping strategies.

The CSQ (20) is a measure of coping strategies for managing pain that is widely used and well validated. It has been extensively used in research on RA (17).

Statistical analysis.

Hierarchical multiple regression analyses with simultaneous entry of variables were used to analyze the data. This method consisted of 2 steps. First, all independent variables, except appearance, were entered to determine the degree to which these accounted for variance in anxiety and depression. For RA, the following variables were entered at stage 1: duration of illness, living arrangements, level of education, age, employment situation, ESR, CRP level, coping strategies, joint inflammation, and disability. For SLE, duration of illness, living arrangements, level of education, age, employment situation, and physical aspects of HRQOL were entered. Second, as suggested by the literature (21), appearance was added to determine whether this mediated the relationship between physical disability (or physical HRQOL for SLE) and distress. Four multiple regression equations were constructed to examine the relationship between appearance and depression and appearance and anxiety. Both patients with recently diagnosed RA and those with chronic RA were combined in 1 set of analyses, and analyses were conducted separately for the SLE group. The RA samples were combined because similar proportions of patients reported appearance concerns in each sample and correlations between variables indicated similar relationships in both samples.


Demographic and clinical characteristics.

The final sample included 60 patients with SLE, 53 with recent-onset RA, and 44 with chronic RA. The demographic and clinical features of these groups are presented in Table 1. The groups were comparable in sex, living arrangements, level of education, and employment. The SLE group was younger than the recent-onset RA and chronic RA groups (mean ± SD age 44.4 ± 12.15 years, 55.06 ± 14.07 years, and 58.63 ± 11.89 years, respectively). All groups showed increased psychological distress as indicated by mean scores on the depression and anxiety subscales of the HADS. The HRQOL of the patients with SLE was negatively affected, with mean ± SD scores for the SF-36 PCS and MCS of 45.6 ± 11.4 and 53.4 ± 11.8, respectively.

Table 1. Demographic and disease variables*
MeasuresSLE groupRecent RA groupChronic RA group
  • *

    Values are the mean ± SD unless otherwise indicated. SLE = systemic lupus erythematosus; RA = rheumatoid arthritis; NA = not applicable; HAQ = Health Assessment Questionnaire; SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; HADS = Hospital Anxiety and Depression Scales.

  • Differences between the groups are significant at P < 0.05.

Female sex, %977075
Age, years44.4 ± 12.1555.06 ± 14.0758.63 ± 11.89
Duration, years10.19 ± 8.691.08 ± 0.8215.5 ± 12.8
Married/cohabiting, %656756
<12 years of education, %424747
Appearance concerns, %533430
In current flare, %55NANA
C-reactive protein levelNA21.93 ± 28.2816.34 ± 21.7
Erythrocyte sedimentation rateNA26.96 ± 18.1422.4 ± 23.0
HAQ (disability)NA0.7 ± 0.580.54 ± 0.54
SF-36 physical composite score45.6 ± 11.4NANA
HADS, depression5.73 ± 3.775.3 ± 3.65.63 ± 3.12
 Rate of likely clinical disorder, %361415
HADS, anxiety7.92 ± 3.047.9 ± 5.17.38 ± 4.69
 Rate of likely clinical disorder, %444134

With regard to concerns about appearance, 53% of the patients with SLE reported that they felt unattractive due to the disease. In contrast, only 30% of the chronic RA sample reported feeling unattractive, which was very similar to the proportion of patients with recently diagnosed RA (34%) who were concerned with their physical appearance.


A series of multiple regression analyses were conducted. The regression models for depression are presented in Table 2 (RA) and Table 3 (SLE). In the combined RA sample, the first step of the multiple regression equation was significant (F[11,86] = 3.457, P = 0.001) and accounted for 24% of the variance in depression. Disability was the only independent predictor of depression in the first step of the equation (t[1,73] = 4.895, P = 0.000). However, when appearance was added, it was significantly associated with depression and added ∼6% to the variance (t[1,85] = 2.823, P = 0.006). However, the relationship between disability and depression was still significant (t[1,86] = 4.523, P = 0.000) (Table 2).

Table 2. Multiple regression equation predicting depression in the rheumatoid arthritis sample*
ModelAdjusted R2βdfF changetP
  • *

    Statistics are presented for the entire model at steps 1 and 2. Only the individual predictors that contributed independently to the variance within each model are reported. HAQ = Health Assessment Questionnaire.

  • Individual predictors: duration of illness, C-reactive protein level, coping strategies, living arrangements, level of education, disability, age, erythrocyte sedimentation rate, employment situation, and joint inflammation.

Step 10.243 11,863.457 0.001
 Significant predictor(s)      
  HAQ: step 1 0.1651,86 4.8950.000
Step 2: add appearance0.309 12,857.972 0.006
 Significant predictors      
  HAQ: step 2 0.1481,85 4.5230.000
  Appearance: step 2 0.2511,85 2.8230.006
Table 3. Multiple regression equations predicting depression in the SLE sample*
ModelAdjusted R2βdfF changetP
  • *

    Statistics are presented for the entire model at steps 1 and 2. Only the individual predictors that contributed independently to the variance within each model are reported. SLE = systemic lupus erythematosus; SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; PCS = physical composite score.

  • Individual predictors: duration of illness, living arrangements, level of education, disease status, sex, employment situation, age, and physical composite score of health-related quality of life.

Step 10.329 8,523.575 0.004
 Significant predictor(s)      
  SF-36 PCS: step 1 −0.1261,52 −3.3950.002
Step 2: add appearance0.513 9,5113.805 0.001
 Individual predictors      
  SF-36 PCS: step 2 −0.06031,51 −1.6570.107
  Appearance: step 2 0.4011,51 3.7160.001

For anxiety, although step 1 of the multiple regression equation was significant (F[11,86] = 2.959, P = 0.003), appearance did not predict anxiety (F[12,85] = 0.000, P = 0.999). Level of disability (t[1,86] = 2.603, P = 0.001) and joint function (t[1,86] = 2.670, P = 0.009) were significantly predictive of anxiety.

Similar relationships were found in the SLE sample. Step 1 was again significant, accounting for 32% of the variance in depression (F[8,52] = 3.575, P = 0.004), and physical HRQOL was the only significant predictor of depression (t[1,52] = 13.805, P = 0.001). However, when appearance was included in the analysis, physical HRQOL was no longer significant (t[1,52] = −1.657, P > 0.1). Appearance, as predicted, was a significant independent predictor for depressive symptoms (t[1,51] = 3.716, P = 0.001), adding 18% to the variance. These results suggest that appearance mediated the relationship between physical HRQOL and depression. As in RA, appearance was not related to anxiety symptoms in SLE (adjusted R2 = 0.160, P > 0.05). Indeed, none of the variables in the regression equation predicted anxiety in the SLE sample (Table 2).


The present study extends the work by Benrud-Larson and colleagues (5) from patients with scleroderma to patients with 2 other rheumatologic disorders. Our results largely supported our hypotheses. That is, we found that both appearance and physical disability are predictive of depression but not anxiety in patients with RA, suggesting that appearance concerns are important and related to depression independently of disability. In patients with SLE, appearance concerns were also related to the outcome of physical HRQOL. A larger proportion of patients with SLE endorsed appearance concerns, and the results demonstrated that appearance concerns mediated the relationship between physical HRQOL and depression.

Participants in each sample frequently reported that their physical appearance had changed due to their disease, with 30–53% of participants reporting that they felt unattractive. This finding suggests that appearance concerns are frequently associated with rheumatic diseases. However, the fact that a similar proportion of patients with recent-onset RA and chronic RA reported concerns regarding physical appearance implies that objective disfigurement is not the sole cause of body image and appearance dissatisfaction in patients with rheumatic disease. Participants in the recently diagnosed group were still in the early stages of the disease and had few, if any, objectively observable disfigurements, but had rates of concerns similar to those with chronic RA.

The importance of these concerns for individuals with rheumatic illness is confirmed by the strong relationship with depressed mood found in this study. We found in separate analyses that appearance was predictive of depression, but not anxiety, in patients with RA and SLE. Individuals with concerns regarding their physical appearance reported more symptoms of depression. This result is similar to those found in studies of other rheumatic diseases (5) and confirms the link between appearance concerns and depression in patients with rheumatic disease.

Disability was also predictive of depression in patients with RA. These results confirm the relationship between physical disability and depressive symptoms in RA (22, 23). Both appearance and disability predicted depression in RA and did so independently. This finding is important because it clarifies that levels of disability do not mediate the relationship between appearance and depression in RA. Therefore, in patients with RA, not all appearance concerns are likely to be manifestations of physical disability and damage associated with illness.

In SLE, appearance mediated the relationship between physical HRQOL and depression. When appearance concerns were not considered, physical HRQOL predicted depression. However, when appearance was added to the model, this was no longer the case. This finding demonstrates that physical limitations associated with the disease predict appearance concerns, which, in turn, predict depression. Therefore, appearance concerns in patients with SLE are largely related to individuals' ability to perform their physical role functions. However, it is the degree to which patients, when faced with these impairments, express concerns about their appearance that is associated with depressed mood. This is consistent with theories of depression in the context of illness, which suggest that it is the way in which illness affects the view of the self that determines the degree of mood disturbance (24).

Although the present results confirm that anxiety levels are elevated in patients with rheumatic diseases, anxiety was not associated with appearance concerns. The only predictor of anxiety was the number of actively inflamed joints in the RA sample. This finding suggests that anxiety may be more strongly related to severity of disease. Nevertheless, although anxiety may be common in rheumatic diseases, appearance concerns are not associated with increased anxiety.

Our results clearly establish that appearance concerns are associated with depression in patients with rheumatic diseases. Although body image dissatisfaction is predictive of depression, this does not establish a causal relationship. It is possible that depression causes body image dissatisfaction, rather than dissatisfaction causing depression. To date, there is no prospective research to confirm the direction of causality. It is likely that the relationship is reciprocal, that is, depression resulting from the disease increases appearance concerns and dissatisfaction intensifies feelings of depression. It is not possible to determine from this study the direction of the relationship between appearance concerns and depression.

Several limitations need to be considered when interpreting the data. First, the recent-onset RA sample was recruited from the UK, whereas the other samples were recruited in Australia. This may have caused differences due to cultural factors. However, the UK and Australia are both Western countries and are typically assumed to be comparable. Therefore, it seems unlikely that differences resulted due to cultural factors. Recruitment also varied between groups. Both RA groups were consecutively recruited from rheumatology clinics, whereas most SLE participants were a convenience sample. This may explain the higher rates of depression and appearance concerns for SLE, rather than illness-related differences. Although this is of concern in interpreting between-group differences, the relationships between variables should be less affected.

Another limitation of this study was the utilization of different measures in the 2 samples. Although physical HRQOL and disability are strongly related (19), they measure different constructs. This may account for the observed differences between RA and SLE. Furthermore, the additional measures used for RA may have caused appearance concerns to appear less salient in RA because more variance was subsumed by other factors. This explanation is unlikely, however, because the patient and illness characteristics in the SLE regression accounted for more of the variance in depression (33%) than in the RA sample (24%). Nonetheless, this explanation cannot be excluded.

The use of different measures produces some problems in comparing the results. However, this was unavoidable because numerous measures used for RA are either unvalidated or irrelevant for SLE. Although this is problematic in drawing direct comparisons, the fact that appearance predicted depression in both samples, despite using different measures, allows more confidence that appearance concerns are important to patients with rheumatic diseases and are related to depressive symptoms. In particular, the regression equations included more than a 1:10 variable-to-case ratio, which is ideal and therefore reduces the power available in these analyses, suggesting a large effect size in both samples (25).

A final limitation of this study was the use of a single-item measure of appearance. Although the DRP has been validated in RA (12) and SLE samples (26), a single-item measure is less robust than a more comprehensive one. This is largely unavoidable because there are no well-validated measures of appearance that have been extensively used with patients with rheumatic diseases. However, future research could work to develop a rheumatology-specific measure of appearance concerns.

These limitations notwithstanding, this study demonstrates that not only are appearance concerns salient for a sizeable proportion of rheumatology patients, but where these concerns are present, they are related to depressive symptoms. Given that depression has been found to predict future disability (11), any measure that can be taken to lessen the distressing impact of the disease on mood should be implemented. Because there is a relationship between appearance concerns and depression, it seems likely that targeting appearance concerns may improve mood.

Cognitive behavioral therapy (CBT) has been shown to reduce depression in patients with RA (13, 20, 27, 28). This type of therapy aims to help patients develop more adaptive attitudes towards the disease and engage in behaviors that will improve their functioning. By expanding CBT to focus on a positive perception of appearance, appearance concerns should decrease. This, in turn, may enhance the effectiveness of CBT in reducing depression. If the importance of appearance to an individual's self-schema can be reduced from the outset, body image dissatisfaction may be reduced or avoided.

In all 3 samples, there were high rates of concerns about appearance. These seemed largely independent of physical factors in the RA group and were strongly associated with depression in both the SLE and RA group. This finding suggests that the assessment of concerns regarding appearance should form part of the routine treatment of patients with rheumatic diseases, even in the absence of objective disfigurement. Where concerns exist, they should be targeted using psychological treatments that have been shown to be effective, such as CBT. What this study highlights is that appearance concerns and depression are closely interrelated and that it is important that they be considered together in the psychological functioning of patients with rheumatic disease.


Dr. Sharpe had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Monaghan, Sharpe, Denton, Sensky.

Acquisition of data. Sharpe, Denton, Levy.

Analysis and interpretation of data. Monaghan, Sharpe, Schrieber, Sensky.

Manuscript preparation. Monaghan, Sharpe, Denton, Levy, Schrieber, Sensky.

Statistical analysis. Monaghan, Sharpe.

Acquisition of research funding. Sharpe.