Anxiety in rheumatoid arthritis

Authors


  • The opinions of this publication are those of the grantee and do not reflect those of the Department of Education or the Department of Veterans' Affairs.

Abstract

Objective

To examine the level of anxiety experienced by individuals with rheumatoid arthritis (RA).

Methods

Data from 2 previous studies were used to compare the level of anxiety (measured by the State-Trait Anxiety Inventory) in the following 4 subgroups: a general RA sample, a general osteoarthritis sample, a sample with both RA and major depression, and a normative sample of age-equivalent, working adults. Canonical correlations were used to examine associations between measures of anxiety and measures of both stress and depression. The relationship between anxiety and duration of RA was also explored.

Results

The general RA sample had state anxiety levels that were comparable to the normative sample, although trait anxiety levels were significantly higher (P < 0.001). In addition, individuals with RA who also met criteria for depression exhibited significantly higher levels of both state anxiety (P < 0.0001) and trait anxiety (P < 0.0001) than was observed in the normative sample. Canonical correlations revealed that measures of anxiety were correlated with both measures of depression (r = 0.83) and measures of stress (r = 0.50). Anxiety was not found to be significantly related to RA disease duration.

Conclusion

These findings demonstrated that individuals with RA, especially if concomitantly depressed, tend to exhibit levels of anxiety that are generally higher than a normative group of age-equivalent, working adults. The substantial canonical correlations between anxiety and both depression and stress revealed that anxiety shares variance with these more frequently studied variables in RA. However, anxiety was not found to be related to RA disease duration.

INTRODUCTION

Anxiety is a common, costly, and serious public health problem. The National Institute of Mental Health Epidemiologic Catchment Area study found a lifetime rate of 14.6% for anxiety disorders (1); and the National Comorbidity Survey found a lifetime rate of 24.9% (2). In 1990, the costs associated with anxiety disorders (e.g., reduced employment and increased medical treatment) amounted to $46.6 billion, which was over 30% of the total expenditures for mental health care that year (3). Mendlowicz and Stein (4) reviewed the literature on quality of life in anxiety disorders; they cited impairment in occupational functioning, social functioning, educational functioning, physical health, mental health, and social support.

Although the importance of psychological adjustment (e.g., stress and depression) has been extensively studied in rheumatoid arthritis (RA) (5–7), anxiety has been largely overlooked. Yet, anxiety is a relevant area to examine; a recent prospective, longitudinal community study revealed that anxiety disorders are often primary conditions (8). Accordingly, the construct of anxiety may be quite important in the context of RA. The symptoms of RA can vary widely over the course of a single day or over longer periods of time, which may contribute to symptoms of anxiety (9). In addition, unpredictable disease course, chronic pain, and movement limitations can all increase the probability of clinically significant affective states, such as anxiety (9, 10).

Hawley and Wolfe (11) noted that anxiety is common in individuals with RA, but found that the level of anxiety in RA was similar to that of other rheumatic disorders (i.e., osteoarthritis, low back pain, and fibromyalgia). Approximately 20% of individuals with RA have been found to exhibit a significant degree of anxiety (10). In a study of Japanese women with RA, over one-third demonstrated a high anxiety level; these women had scores on the State-Trait Anxiety Inventory (STAI) that were approximately one standard deviation above the mean for female working adults (12). An American study found elevated trait anxiety scores for persons with RA (13), but other studies have reported average levels of anxiety in an RA population. A German study found that persons with RA exhibited average trait anxiety scores, which were similar to the normative data for the general population (14). Similarly, Anderson et al (15) reported average STAI scores for persons with RA. These conflicting results regarding the prevalence of anxiety in RA warrant further investigation.

The purpose of this study was to address 3 specific research questions. Do anxiety levels for various arthritis subgroups differ from anxiety levels reported by a normative group? Is anxiety a useful measure of psychological distress in RA beyond the constructs of depression and stress? Is anxiety level correlated with disease duration in RA?

METHODS

Participants.

The current analyses were conducted on baseline data from previously published arthritis intervention studies. The first 2 studies evaluated the effects of stress management on general samples of individuals with RA (16) and osteoarthritis (OA) (17). The third study evaluated the effectiveness of depression management in a sample of individuals with concomitant major depression and RA (18). Participants in all 3 studies were recruited from a Midwestern Department of Veterans' Affairs hospital, a university medical center, and a private rheumatology clinic. There were 2 samples: a general sample of persons with RA and OA who were selected to be as representative as possible, and a depressed sample that met the criteria for major depression.

The general sample consisted of individuals with RA (n = 436) who were screened for exclusion criteria for a stress-management intervention. Individuals with psychotic disorders, other uncontrolled medical disorders, Functional Class IV (19, 20), or a history of organic brain disorder were excluded. The general sample, which was selected for the stress-management protocol, included 143 people with RA (81 male, 62 female), and 29 people with OA (14 male, 15 female). The mean ± SD age for the RA sample was 58.6 ± 10.7 years, and 59.5 ± 8.3 years for the OA sample. The mean ± SD disease duration for persons with RA and OA was 12.2 ± 9.7 years, and 8.8 ± 8.8) years, respectively. A more detailed description of these samples can be found in Parker et al (16) and Wright et al (17).

The depressed sample consisted of individuals with RA (n = 638) who were screened for depression. Those who scored 12 or higher (n = 254) on the Center for Epidemiologic Studies-Depression Scale (CES-D) were invited to complete a structured diagnostic interview. Eighty-four participants consented, and of these, 54 (15 male, 39 female) were diagnosed with major depression. The mean ± SD age was 54.7 ± 11.4 years, and RA disease duration was 12.6 ± 10.1 years. A more detailed description of this sample can be found in Parker et al (18).

Existing normative data on the STAI was used to create an age-equivalent comparison group of working adults (21).

Measures.

The measures used were the STAI, the Symptom Checklist 90-R (SCL-90-R), the Arthritis Impact Measurement Scales (AIMS), the CES-D, the Daily Stress Inventory (DSI), and the Hassles Scale (HS). The STAI is a 40-item measure used to assess state and trait anxiety. Both the state and trait anxiety subscales from the STAI have been found to be reliable and valid (21). The STAI has been used successfully in previous research with people with RA (15). In the current study, Cronbach's coefficient alpha for the 2 STAI subscales ranged from 0.89 to 0.90.

The SCL-90-R is a 90-item, self-report inventory designed to measure psychological distress. It has been shown to have excellent reliability and validity (22). In the current study, the depression and anxiety subscales were used, and Cronbach's coefficient alphas were 0.80 and 0.72, respectively.

The AIMS is a 66-item questionnaire designed to measure health status for persons with arthritis (23). Adequate reliability and validity data for the AIMS have been reported (23, 24). In the current study, the depression and anxiety subscales were used, and Cronbach's coefficient alphas were 0.84 and 0.90, respectively.

The CES-D is a 20-item self-report scale that measures depressive symptoms (during the past week) in the general population (25). The CES-D has high internal consistency, adequate test-retest reliability, and good factorial and discriminant validity (26). There is evidence that the CES-D is a valid measure of depression for individuals with arthritis (27). In the current study, Cronbach's coefficient alpha was 0.87.

The DSI is a 15-item measure of stressful events in daily life (during the past 24 hours). The DSI has good internal consistency and validity (28). In the current study, Cronbach's coefficient alpha was 0.91 for the DSI (severity of stressors score); the Kuder-Richardson 20 for the DSI (frequency of stressors score) was 0.90.

The HS is a 117-item measure of general life stress, which quantifies pressures, problems, and difficulties of everyday life (during the past month). Adequate reliability and validity for the HS have been reported (29). In the current study, Cronbach's coefficient alpha was 0.94 for the severity of daily stressors score; the Kuder-Richardson 20 was 0.93 for the number of daily stressors score.

Procedures.

In both samples, only baseline measures were analyzed to prevent contamination by the intervention effects that were under examination in the original studies (16–18). The participants in the general sample (both RA and OA) completed the STAI (state and trait anxiety subscales), the SCL-90-R (anxiety and depression subscales), the AIMS (anxiety and depression subscales), the CES-D, the DSI, and the HS. Participants in the depressed RA sample completed the STAI (state and trait anxiety subscales).

Data analyses.

Analyses were performed in the following four phases: descriptive statistics were generated for the STAI for both the general sample (RA and OA) and the depressed sample (RA) and the STAI scores were compared with a normative group of age-equivalent, working adults; canonical correlations were used to examine the relationship between anxiety measures and depression measures; canonical correlations were used to examine the relationship between anxiety measures and stress measures, and finally, correlations between level of anxiety (state and trait) and duration of illness were examined in the general sample of individuals with RA. For the group comparisons, each of the 3 arthritis samples were compared with the normative sample on measures of both state and trait anxiety (which resulted in a total of 6 comparisons). Therefore, a Bonferroni correction for multiple comparisons was used; α = 0.008 (0.05/6).

RESULTS

Group comparisons on the STAI.

The pretreatment STAI means and standard deviations were used to compare the level of anxiety in people with arthritis with the data reported in the STAI normative group (21). The groups were compared as follows: a general RA sample, a general OA sample, a sample with both RA and major depression, and data from the normative sample of age-equivalent (range 50–69 years old), working adults. Paired t-tests revealed significantly higher state anxiety scores for the depressed RA sample in comparison with the normative group (P < 0.0001), but no significant difference in state anxiety was found between either the general RA sample or the general OA sample in comparison with the normative group (Table 1). Paired t-tests revealed significantly higher trait anxiety scores for both the general RA sample (P < 0.001) and the depressed RA sample (P < 0.0001) in comparison with the normative group, but no significant difference in trait anxiety was found between the general OA sample and the normative group (see Table 1).

Table 1. Statistics for State-Trait Anxiety Inventory in RA, OA, and normative samples*
SampleFirst author (reference)State anxietyTrait anxiety
  • a

    Values are mean ± SD. RA = Rheumatoid arthritis; OA = osteoarthritis.

  • b

    P < 0.001 versus normative sample, on trait anxiety scores.

  • c

    P < 0.0001 versus normative sample on state anxiety, and trait anxiety scores, respectively.

  • *, †, ‡, §

    Normative sample = working adults ages 50–69 years.

General RAParker (16)33.06 ± 8.6136.17 ± 9.22
General OAWright (17)31.93 ± 8.5333.72 ± 8.80
Depressed RAParker (18)46.83 ± 10.2448.22 ± 10.38
Normative§Speilberger (21)34.01 ± 10.0033.41 ± 8.64

Anxiety and depression.

Canonical correlation is a multivariate statistical technique that summarizes the relationship between 2 sets of variables by finding linear combinations that maximize the correlation between the 2 sets. Canonical correlations were performed to investigate the relationship between measures of anxiety (STAI state and trait, SCL-90-R anxiety, AIMS anxiety) and measures of depression (CES-D, SCL-90-R depression, AIMS depression) in the general RA sample. The canonical correlation between the anxiety and the depression measures revealed a statistically significant relationship (r = 0.83, R2 = 0.69).

Anxiety and stress.

Canonical correlations also were performed to investigate the relationship between measures of anxiety and stress in the general RA sample; the canonical correlation between anxiety measures (STAI state and trait, SCL-90-R anxiety, AIMS anxiety) and stress measures (DSI frequency and severity, HS number and severity) revealed a statistically significant relationship (r = 0.55, R2 = 0.30).

Anxiety and disease duration.

The relationship between level of anxiety and disease duration was examined in the general RA sample. The correlations between level of anxiety (as measured by STAI state and trait subscales) and number of months since being diagnosed with RA were not statistically significant.

DISCUSSION

This study yielded important findings about anxiety and RA, a topic that has been relatively overlooked in the research literature. First, the general RA sample exhibited state anxiety scores that were comparable to those found in the normative group, but the trait anxiety scores for the general RA sample were significantly higher than those for the normative group. RA is a challenging and severe illness, so an elevated anxiety score is not entirely surprising; numerous conditions associated with RA (e.g., pain, decreased functional status, reduced income) may contribute to the observed elevation in trait anxiety.

A second finding was that individuals with RA who were identified as being depressed exhibited notably high levels of anxiety. The canonical correlations between anxiety and depression indicated that depression and anxiety have a substantial overlap (sharing 69% of the variance). These results support previous findings by Hagglund et al (30), who found that measures of both depression and anxiety assess general distress. Also, Clark and Watson's (31) tripartite model (i.e., depression, anxiety, and distress) theorized that the shared factor of distress accounts for the overlap between anxiety and depression. Not surprisingly, anxiety and depression are commonly comorbid conditions and are typically correlated (32, 33). In addition, previous research has revealed that a frequent temporal sequence is anxiety first and then subsequently depression (8, 34). Therefore, screening for symptoms of anxiety in persons with RA might facilitate early identification of depression and, thereby, help to prevent future depressive episodes.

A third finding was that measures of anxiety and stress have substantial overlap (sharing 25% of variance). Anxiety and stress are related concepts, so shared variance was expected. However, the analyses highlighted the importance of the distinction between stress (measured by the number and severity of everyday stressful events) and anxiety (measured by the experience of an emotional state); measures of both anxiety and stress exhibited substantial unique variance.

A fourth finding was that level of anxiety was not related (in linear fashion) to disease duration in RA. Over the course of their disease, persons with RA showed a wide range of anxiety levels. This finding may appear to be somewhat inconsistent with previous reports showing slight decreases in anxiety over the first few years of having RA (11, 15, 35), but the current sample was comprised of individuals with a relatively long disease duration. In addition, the association between anxiety and disease duration may ultimately prove to be curvilinear.

These analyses were limited by the absence of a diagnostic evaluation of anxiety disorders; only data from general screening questionnaires were available. The non-random sample also limits generalizability to some extent. However, these findings suggest that persons with RA exhibit levels of trait anxiety that are higher than those observed in an age-equivalent, normative group and that persons with both RA and major depression exhibit much higher levels of anxiety (both state and trait) than does an age-equivalent, normative group. Given that anxiety can be a precursor for subsequent depression, screening and early detection in rheumatology settings appears to be a worthwhile consideration.

Ancillary