Depression, fatigue, and pain in systemic lupus erythematosus (SLE): Relationship to the American College of Rheumatology SLE neuropsychological battery

Abstract

Objective

To examine the frequency and reliability of depression, fatigue, and pain self-report measures in patients with systemic lupus erythematosus (SLE) and healthy controls, and to examine the relationship between a cognitive impairment index (CII) derived from the American College of Rheumatology neuropsychology research battery of tests for SLE (ACR-SLE battery) and measures of depression, pain, fatigue, and perceived cognitive dysfunction.

Methods

Thirty-one patients with SLE with a history of overt neuropsychiatric symptoms (neuropsychiatric SLE [NPSLE]), 22 patients with SLE without overt neuropsychiatric symptoms (non-NPSLE), and 25 healthy controls completed the following measures at baseline and 1-month followup: ACR-SLE battery, perceived cognitive difficulties, depression, fatigue, and pain.

Results

Patients with SLE (both NPSLE and non-NPSLE) showed higher symptoms of depression, higher levels of fatigue, greater pain, and more perceived cognitive problems. All measures except the Center for Epidemiologic Studies Depression scale (CES-D) demonstrated adequate reliability across the SLE groups at retest. Only patients with NPSLE had significant correlations between CII and depression, fatigue, and pain. Neither the non-NPSLE patients nor the controls had significant relationships with the CII and these behavioral measures.

Conclusion

Patients with SLE report higher levels of cognitive difficulties, depression, pain, and fatigue compared with controls. Reliability for all measures, except the CES-D, was established in the SLE group. Overall, results suggest that cognitive dysfunction, pain, fatigue, and depression in patients with NPSLE may represent global changes in the central nervous system that require ongoing evaluation and treatment.

INTRODUCTION

Standard nomenclature for 19 neuropsychiatric systemic lupus erythematosus (SLE) syndromes was developed by the American College of Rheumatology (ACR) (1). Cognitive dysfunction was identified as one of the neuropsychiatric syndromes for SLE and was defined as “significant deficits in any or all of the following cognitive functions: complex attention, executive skills (e.g., planning, organizing, sequencing), memory (e.g., learning, recall), visual-spatial processing, language (e.g., verbal fluency), and psychomotor speed” (1). The ACR committee proposed a standard battery of neuropsychological tests for SLE (referred to as the ACR-SLE battery). In our recent study (2), we established the reliability and validity of this battery of tests. The ad hoc ACR committee also identified associations that could impact cognitive test scores in patients with SLE (1), including depression, fatigue, and pain. Specific measures were suggested and future studies in relation to cognitive performance on the ACR-SLE battery were strongly suggested.

Depression and emotional distress have been reported in ∼50% of patients with SLE (3). There is considerable debate regarding the prevalence and the diagnosis of depression in patients with SLE (4), and methodologic problems (i.e., subject selection, measurement) have limited conclusions across studies. The relationship between cognitive dysfunction and depression in SLE has varied. Most commonly, patients with overt neuropsychiatric disorders have shown higher associations between cognitive impairment and depression than patients without overt neuropsychiatric disorders (5–7). In patients with SLE without overt neuropsychiatric SLE (NPSLE), or in studies primarily of patients with non-NPSLE, no associations between depression and cognition have been reported (8, 9). The Center for Epidemiologic Studies Depression (CES-D) (10) measure is a common self-report instrument of depression, with high levels reported in 35–85% of patients with SLE (11–13). This scale was selected by the ACR committee as a potential measure of depression in SLE.

Fatigue has been reported in 50–100% of patients with SLE and has been described as the most chronic symptom these patients experience (12–18). Many patients report not having enough energy to perform daily activities (15) and that fatigue frequently compromises activities, affects physical functioning, interferes with work, and reduces quality of life (12, 13). Two studies have reported an association between fatigue and neurologic manifestations in SLE (17, 19). Gilboe et al (20) reported that fatigue was more strongly associated with neuropsychiatric symptoms in comparison with musculoskeletal and pulmonary symptoms. However, fatigue was not found to correlate with overt central nervous system symptoms or cranial magnetic resonance imaging results in SLE (21). To date, few studies have correlated fatigue with neuropsychological functioning in SLE. In a pilot study of 10 patients with SLE, Denburg et al (22) did not find a relationship between fatigue and cognition. However, Alarcon et al (23) report that self-reported cognitive difficulties are related to fatigue. The Fatigue Severity Scale (FSS) (12) and the Multidimensional Assessment of Fatigue (MAF) (24) were utilized in a variety of autoimmune studies, and were utilized in the present study to measure aspects of fatigue.

Although pain has been reported in patients with SLE (25), few studies have investigated pain in SLE in relation to cognition. Denburg et al (22) found that in 10 patients with SLE there was an association between pain (measured by visual analog scale) and overall cognitive dysfunction, particularly learning and memory. Self-reported cognitive problems were also higher in patients with SLE with greater pain (23). There is also evidence that pain may reduce overall cognitive efficiency in other populations (26, 27). The Short-Form McGill Pain Questionnaire (SF-MPQ) has been used in a variety of studies of pain in other clinical populations (28), and was therefore selected for this study.

Self-report measures of cognitive functioning in SLE have also been recommended by the ACR committee for clinical and research purposes. As recommended, the Cognitive Failures Questionnaire (CFQ), a brief measure of self-reported cognitive problems, was included (29). The Patient's Assessment of Own Functioning Inventory (PAF) (30), which has been utilized in prior studies, was also administered as a subjective measure of neurocognitive symptoms.

This study aimed first to compare the performances of patients with NPSLE, patients with non-NPSLE, and controls on various measures of depression, fatigue, pain, and perceived cognitive problems. Next, we aimed to examine the reliability of these tests after a 1-month followup. Finally, we planned to examine associations between performance on the ACR-SLE battery (total cognitive impairment index) and measures of depression, fatigue, and pain.

PATIENTS AND METHODS

Patients.

Participants in this study included 31 patients with SLE with a history of neuropsychiatric disorders (NPSLE), 22 patients with non-NPSLE, and 25 healthy controls. All participants signed an approved consent form authorized by the Institutional Review Board at the National Jewish Medical and Research Center (Denver, CO). The SLE patients were obtained from a pool of patients with SLE seen at the National Jewish Medical and Research Center, The University of Colorado Health Sciences Center, and local rheumatology clinics. Patients with SLE with possible neurologic damage prior to SLE (head trauma, degenerative, vascular or metabolic disorder, neoplasm, or toxic exposure), history of major substance abuse, or major psychopathology were excluded from the study. All patients with SLE fulfilled the revised criteria for SLE as defined by the ACR (31). Details regarding recruitment and evaluation procedures have been reported previously (2). A checklist containing the 19 neuropsychiatric syndromes described by the ACR was completed by the physician of each patient with SLE to classify patients as NPSLE or non-NPSLE. Our rheumatology consultant (SW) reviewed all checklists and assured accuracy through contact with physicians. Participants were seen for an initial evaluation and a 1-month followup appointment. None of the participants had major medical or new neuropsychiatric events within this period as self-reported during a followup interview.

Participant demographics and health characteristics can be found in Table 1. There were 20 women and 2 men with NPSLE, 30 women and 1 man with non-NPSLE, and 23 women and 2 men in the healthy control group. As indicated in Table 1, the groups did not significantly differ in age, education level, sex distribution, or race/ethnicity. SLE disease activity was measured with the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) (32), and was obtained from each patient's rheumatologist or primary physician at the time of enrollment. As indicated in Table 1, SLEDAI scores were similar across the 2 SLE groups, and a mean score of 6 suggested mild to moderate disease activity. The SLE groups did not differ in terms of length of disease, with a mean disease duration of 8.5 years in the NPSLE group and 6.4 years in the non-NPSLE group. As previously reported, a majority of the 31 patients with NPSLE had a history of >1 neuropsychiatric syndrome and 48% had mood disorder; 45% headache; 6% stroke, movement disorder, or psychotic episode; and 3% asceptic meningitis (2). At the time patients were studied, 61% of those with NPSLE and 45% of those with non-NPSLE were taking prednisone. The mean dose of prednisone was similar across the NPSLE and non-NPSLE groups (7.28 mg and 6.05 mg, respectively). Additional medications in the NPSLE and non-NPSLE groups at the time of the study included nonsteroidal antiinflammatory drugs (52% versus 36%), antidepressants (48% versus 50%), thyroid medication (61% versus 14%), antihypertensives (45% each), estrogen (13% versus 45%), gastrointestinal medications (45% versus 41%), and opioid analgesics (29% versus 9%). As noted in Table 1, the Cognitive Impairment Index (CII) varied significantly across our groups, and post hoc analysis indicated that the NPSLE group had a higher CII than both the non-NPSLE group (P = 0.029) and the control group (P = 0.003); however, the non-NPSLE group and controls did not differ (P = 0.127).

Table 1. Demographics and health characteristics for patients with NPSLE, non-NPSLE, and healthy controls*

Characteristic	NPSLE (n = 22)	Non-NPSLE (n = 31)	Controls (n = 25)	P
* Values are the mean ± SD unless otherwise indicated. NPSLE = neuropsychiatric systemic lupus erythematosus; NA = not applicable; SLEDAI = Systemic Lupus Erythematosus Disease Activity Index.
Demographics
Age, years	44.8 ± 11.1	44.4 ± 12.6	43.5 ± 11.5	0.88
Education, years	15.2 ± 2.4	14.4 ± 2.7	15.4 ± 2.0	0.31
Sex, female/male	20/2	30/1	23/2	0.61
Ethnicity, no.				0.69
African American	2	2	2
White	18	27	23
Hispanic	2	2	0
Health characteristics
Length of diagnosis	102.3 ± 93.6	76.5 ± 77.6	NA	0.20
SLEDAI	6.6 ± 5.4	6.1 ± 4.8	NA	0.74
Cognitive Impairment Index	3.74 ± 2.65	2.41 ± 1.92	1.44 ± 1.58	0.0007

Measures.

The analyses for this study utilized several cognitive batteries of tests and measures of depression, pain, fatigue, and perceived cognitive ability.

ACR-SLE battery.

The ACR-SLE battery, a battery of tests proposed by the ACR for SLE (1), was administered by a trained neuropsychological technician. The following tests were administered and the following selected scores were used in the analyses: Wechsler Adult Intelligence Scale (WAIS), Revised Digit Symbol Test, total number (33); Trail Making Test-Form B, total time (34); Stroop Color and Word Test-Color-Word, total score (35); California Verbal Learning Test-II-Trials 1–5 Learning, delayed free recall (36); Rey-Osterrieth Complex Figure Test-Learning score, delayed recall score (37); WAIS-III Letter Number Sequencing, total score (38); Controlled Oral Word Association Test and Animal Naming Test, total score (39); and Finger Tapping Test, dominant and nondominant hands (34). Reliability and validity for this battery of tests have been documented (2) via comparison with a larger battery of tests and test–retest analysis. In addition, a CII can be calculated utilizing the 12 selected test scores. Each score was converted to a T score using demographically corrected normative data, and a T score <40 was considered impaired. The CII has a range of 0–12, with a higher number representing greater cognitive impairment.

The Center for Epidemiologic Studies Depression Scale (10).

The CES-D is a self-administered, 20-item questionnaire that measures the individual's state (on a scale of 0 = “rarely or none of the time” to 3 = “most or all of the time”) with regard to mood and vegetative motor functions during the preceding week. CES-D total scores range from 0 to 60. A score ≥16 indicates clinical depression (10). This scale has demonstrated adequate reliability and validity in various settings and across ethnic backgrounds (40).

The Fatigue Severity Scale (12).

The FSS questionnaire is a 9-item scale of fatigue. Individuals indicate the extent to which they agree with each statement of fatigue's impact on activities of daily living on a scale of 1 (“strongly disagree”) to 7 (“strongly agree”). The FSS total score is the mean of these 9 scores. This scale has demonstrated adequate reliability and validity in patients with SLE, multiple sclerosis, and chronic fatigue syndrome (12, 18). Fatigue is strongly associated with psychological dysfunction and emotional disturbances in patients with SLE (12). An FSS total score was calculated for analysis.

Modified Multidimensional Assessment of Fatigue Questionnaire (24).

The MAF, a self-administered, 16-item questionnaire, assesses 3 dimensions of fatigue over the past week: severity and distress (items 1–3), impact on activities of daily living (items 4–14), and timing (items 15 and 16). The first 14 items are rated on a scale of 1 (“not at all”) to 10 (“a great deal”). To calculate the MAF global fatigue index, items 1–3 are added to the average of items 4–14, and the converted score for item 15. Reliability and validity were established for this instrument in older adults with rheumatoid arthritis (41).

The Short-Form McGill Pain Questionnaire (28).

The SF-MPQ provides a qualitative and quantitative assessment of pain. It contains 15 pain-related words divided into sensory and affective categories in a pain-rating index. Individuals are asked to give each description of pain a rating from 0 (“none”) to 3 (“severe”) based on the degree to which they feel that type of pain. Total pain is calculated by adding up these ratings. The range is from 0 to 45 and is labeled McGill Pain Questionnaire total.

Cognitive Failures Questionnaire (29).

This self-administered, 25-item questionnaire measures everyday cognitive errors of attention, perception, memory, and motor functioning over the past 6 months on a 5-point scale that ranges from 0 (“never”) to 4 (“very often”). Total scores for the CFQ range from 0 to 100.

Patient's Assessment of Own Functioning Inventory (30).

The PAF is a subjective neurocognitive symptom questionnaire, with questions in 5 scales: memory, language and communication, use of hands, sensory perception, and executive functions. Individuals answer questions on a scale of 0 (“almost never”) to 5 (“almost always”). The mean score is calculated for a PAF total score, along with subscores for memory, language and communication, use of hands, and executive function.

Statistics.

Data are presented as the mean ± SD for quantitative data and number of participants for qualitative data. A 1-way analysis of variance (ANOVA) and chi-square test were used to evaluate the overall group differences in demographic and health characteristic variables. A 1-way ANOVA was also used to compare measures of depression, fatigue, and pain among the 3 groups. The Kruskal-Wallis test was used for non-Gaussian data. Post hoc analyses were performed using Bonferroni multiple comparisons procedure with the experiment-wise Type I error rate at the 5% level. Intraclass correlation coefficients (r) were calculated via a 1-way random-effects model ANOVA. Generalized linear models with Poisson and binomial distributions were used to evaluate relationships of CII, SLEDAI, or prednisone use with measures of depression, fatigue, and pain. A 2-tailed P value less than 0.05 was considered statistically significant.

RESULTS

Comparison of behavioral measures across groups.

We compared the SLE and control groups on measures of depression, fatigue, and pain. As indicated in Table 2, significant group differences in CES-D total score were found. Post hoc analysis indicated that both patients with NPSLE and those with non-NPSLE had higher CES-D total scores compared with controls. The FSS total score and MAF global fatigue index were significantly higher in the SLE groups compared with controls, and were correlated in both of the SLE groups (P < 0.001) and in the control group (P = 0.04). The McGill Pain Questionnaire total score was significantly higher in both SLE groups compared with controls. The CFQ total score and the PAF total score indicated higher self-reported problems in both SLE groups compared with controls. Both SLE groups had significantly higher PAF memory and language subscores. Both SLE groups also had higher PAF use of hands and executive functions subscores; however, the NPSLE group had higher subscores than the non-NPSLE group in these areas. The PAF and CFQ total scores were highly correlated in each of the groups (P < 0.001).

Table 2. Comparison of depression, fatigue, pain, and perceived cognitive dysfunction across groups*

	NPSLE (1)	Non-NPSLE (2)	Controls (3)	P	Post hoc analyses
* Values are the mean ± SD unless otherwise indicated. NPSLE = neuropsychiatric systemic lupus erythematosus; CES-D = Center for Epidemiologic Studies Depression Scale; FSS = Fatigue Severity Scale; MAF = Multidimensional Assessment of Fatigue; SF-MPQ = Short Form McGill Pain Questionnaire; CFQ = Cognitive Failures Questionnaire; PAF = Patient Assessment of Functioning.
CES-D total score	21.3 ± 7.5	22.8 ± 5.3	14.6 ± 4.6	< 0.0001	1,2 >3
FSS total score	5.4 ± 1.26	4.98 ± 1.3	2.4 ± 0.8	< 0.0001	1,2 >3
MAF global fatigue index	34.7 ± 9.0	29.5 ± 12.1	9.9 ± 8.3	< 0.0001	1,2 >3
SF-MPQ total score	13.5 ± 9.2	11.6 ± 8.0	1.1 ± 2.1	< 0.0001	1,2 >3
CFQ total score	53.0 ± 15.3	42.6 ± 17.3	26.8 ± 10.4	< 0.0001	1 >2; 1,2 >3
PAF total score	1.8 ± 0.8	1.3 ± 0.7	0.5 ± 0.3	< 0.0001	1 >2; 1,2 >3
PAF memory	2.0 ± 0.8	1.7 ± 0.9	0.7 ± 0.4	< 0.0001	1,2 >3
PAF language	1.9 ± 0.8	1.5 ± 0.9	0.7 ± 0.4	< 0.0001	1,2 >3
PAF use of hands	1.6 ± 1.0	1.1 ± 1.0	0.3 ± 0.4	< 0.0001	1 >2; 1,2 >3
PAF executive functions	1.7 ± 1.0	1.1 ± 0.7	0.3 ± 0.3	< 0.0001	1 >2; 1,2 >3

Reliability of measures.

Intraclass correlations performed across measures suggested poor reproducibility (r < 0.4) for the CES-D total score in the non-NPSLE group (Table 3). All other measures in the SLE groups showed fair to good or excellent reliability. Excellent reproducibility (r = 0.86) was found for the FSS total score in the non-NPSLE group. The MAF global fatigue index was excellent for both SLE groups. The CFQ total score, the PAF total score, and the PAF language subscore were excellent for all groups. Additionally, the PAF memory and language subscores were excellent in both SLE groups, and the PAF use of hands and executive functioning subscores were excellent in the NPSLE group.

Table 3. Intraclass correlation coefficients by group*

	NPSLE	Non-NPSLE	Controls
* See Table 2 for definitions. † Reliability coefficients are poor (r < 0.4).
CES-D total score	0.52	0.23†	0.74
FSS total score	0.61	0.86	0.73
MAF global fatigue index	0.77	0.91	0.56
SF-MPQ total score	0.79	0.89	0.61
CFQ total score	0.85	0.83	0.88
PAF total score	0.89	0.83	0.77
PAF memory	0.77	0.78	0.67
PAF language	0.77	0.83	0.76
PAF use of hands	0.78	0.59	0.61
PAF executive functions	0.86	0.56	0.61

Relationship between CII and measures of depression, fatigue, and pain.

A CII was calculated according to procedures previously published (2) and ranged from 0 to 12 (higher score indicating greater cognitive impairment). The patients with NPSLE had significantly higher cognitive impairment compared with those with non-NPSLE and controls (Table 1). Only the CII in the NPSLE group had significant correlations with these measures of depression, fatigue, and pain (Table 4). A higher CII in patients with NPSLE was significantly associated with the CES-D total score (P = 0.026), greater McGill Pain Questionnaire total score (P < 0.0001), and higher FSS total score (P = 0.007) in NPSLE. Additionally, higher MAF global fatigue index was associated with higher CII (P < 0.0001) in NPSLE. Neither patients with non-NPSLE nor the healthy controls showed significant correlations between the CII and any of the depression, pain, and fatigue measures.

Table 4. Generalized linear model of Cognitive Impairment Index, SLEDAI, and prednisone to measures of depression, fatigue, pain, and perceived cognitive functioning*

	Cognitive Impairment Index		SLEDAI		Prednisone
	Parameter estimates	P	Parameter estimates	P	Parameter estimates	P
* SLEDAI = Systemic Lupus Erythematosus Disease Activity Index; see Table 2 for additional definitions.
CES-D total score
NPSLE	0.023	0.026	−0.011	0.211	−0.050	0.271
Non-NPSLE	0.002	0.897	−0.013	0.157	−0.082	0.088
FSS total score
NPSLE	0.017	0.007	0.005	0.297	0.024	0.354
Non-NPSLE	0.006	0.449	0.005	0.372	0.006	0.830
MAF global
NPSLE	0.025	< 0.0001	0.012	0.031	−0.017	0.552
Non-NPSLE	0.010	0.156	0.013	0.028	−0.046	0.154
SF-MPQ total score
NPSLE	0.034	< 0.0001	0.010	0.152	−0.034	0.354
Non-NPSLE	−0.002	0.885	0.022	0.007	−0.087	0.078
CFQ total score
NPSLE	0.014	0.001	0.006	0.065	−0.007	0.693
Non-NPSLE	0.007	0.198	0.008	0.059	−0.017	0.424
PAF total score
NPSLE	0.498	< 0.0001	0.200	0.004	0.074	0.841
Non-NPSLE	0.339	0.009	0.240	0.009	−0.179	0.710
PAF memory
NPSLE	0.408	< 0.0001	0.064	0.335	0.048	0.889
Non-NPSLE	0.218	0.041	0.061	0.427	−0.068	0.866
PAF language
NPSLE	0.476	< 0.0001	0.260	< 0.0001	−0.005	0.987
Non-NPSLE	0.316	0.003	0.227	0.003	−0.411	0.330
PAF use of hands
NPSLE	0.331	< 0.0001	0.034	0.568	0.205	0.500
Non-NPSLE	0.110	0.323	0.167	0.015	0.051	0.894
PAF executive functions
NPSLE	0.365	< 0.0001	0.129	0.024	−0.033	0.912
Non-NPSLE	0.153	0.266	0.164	0.083	−0.431	0.382

The CII was highly associated with perceived cognitive deficits using the CFQ total score in the NPSLE group only (P = 0.001). The PAF total score was highly associated with CII in the NPSLE group (P < 0.0001) and in the non-NPSLE group (P = 0.009). In addition, the PAF memory subscore was highly associated with CII for both the NPSLE (P < 0.0001) and non-NPSLE (P = 0.041) groups, and the PAF language subscore was associated with CII for the NPSLE (P < 0.0001) and non-NPSLE groups (P = 0.003). In the NPSLE group only, PAF use of hands and executive functions were both highly correlated with CII (P < 0.0001).

Relationship of SLE disease activity and prednisone dose to CII and depression, fatigue, and pain measures.

Disease activity in patients with SLE (as measured by the total score on the SLEDAI) was significantly associated with the CII in both the NPSLE (P < 0.001) and non-NPSLE groups (P = 0.018). The SLEDAI was not significantly associated with the CES-D total score in either group. The SLEDAI was highly associated with McGill Pain Questionnaire total for the non-NPSLE group (P = 0.007). The SLEDAI was not associated with the FSS total score, but was associated with the MAF global fatigue index for the NPSLE (P = 0.031) and non-NPSLE groups (P = 0.028). Finally, the SLEDAI was associated with PAF total score for the NPSLE (P = 0.004) and non-NPSLE groups (P = 0.009), but was not associated with the CFQ total score for either group. Prednisone dose was not significantly related to the CII or any of the measures of depression, fatigue, pain, or perceived cognitive difficulties.

DISCUSSION

Consistent with prior studies, higher levels of depression, fatigue, pain, and perceived cognitive problems occurred in both patients with NPSLE and those with non-NPSLE compared with controls (5, 7, 14, 16–18, 42). Both NPSLE and non-NPSLE groups showed similar elevations on all measures except perceived cognitive problems. For this measure, the patients with NPSLE reported higher levels compared with patients with non-NPSLE, a finding that accurately reflects greater objective cognitive deficits in NPSLE as documented by a higher CII.

A majority of the behavioral measures (several recommended by the ACR ad hoc committee) demonstrated adequate reliability in a test–retest design (FSS, MAF, SF-MPQ, PAF, CFQ) and should be considered in future clinical applications and empirical designs. In contrast to the other measures, the CES-D did not demonstrate adequate reliability at retest in the non-NPSLE group. Although the sensitivity and specificity of this measure in SLE and autoimmune disorders have been noted previously (43, 44), our findings would suggest that some caution should be exercised when using this measure in patients with SLE. The reliability coefficient in another study of the CES-D with patients with autoimmune diseases was higher using intertest reliability coefficients (45). The test–retest reliability coefficient for the Beck Depression Inventory was 0.57 for patients with SLE who completed a mailed survey and 0.77 for patients with non-NPSLE who were tested in the laboratory (46), suggesting the latter as a preferred measure. However, because changes in clinical disease and psychological status were measured subjectively versus objectively in our study, it is possible that subtle clinical changes occurred and influenced the test–retest outcome for the CES-D in the non-NPSLE group. To rule out any clinical variability and establish error measurement of the instrument, future studies should include a lupus disease index and structured psychological interviews at followup.

The NPSLE group showed strong associations between cognitive impairment and depression, pain, and fatigue. This finding is not surprising given the fact that many of the patients with NPSLE in this study had mood disorders and headaches, and prior studies have reported relationships between cognition and depression in patients with NPSLE (5–7). Although associations between these factors are strong in NPSLE, the underlying causal factors are unclear. Current research indicates that lupus antibodies cross-react with both DNA and N-methyl-D-aspartate (NMDA) receptors in SLE (47), and elevated levels of antibodies against NMDA receptors have been associated with depressed mood and learning deficits in persons with NPSLE (48). It is therefore possible that associations between cognition and behavioral factors such as depression, pain, and fatigue may reflect an underlying central nervous system disorder. However, disease activity and inflammation may also contribute to behavioral changes. Future studies will require larger sample sizes and a variety of biologic measures (i.e., immune activity, brain imaging, etc.) to better model these relationships.

Neither the patients with non-NPSLE nor the controls had relationships between fatigue, pain, or depression and overall cognitive dysfunction with cognitive impairment. These results are consistent with prior studies that report that depression is not related to cognition in non-NPSLE (8, 9). This finding is important to note and would suggest that cognitive dysfunction identified in patients with non-NPSLE remains an important independent neuropsychiatric syndrome. Continued evaluation of biologic mechanisms associated with cognitive dysfunction in this particular group will facilitate our understanding of this specific neurocognitive dysfunction in SLE.

Patients with SLE had higher self-reported cognitive symptoms compared with controls, and the symptoms of those with NPSLE were higher than those of patients with non-NPSLE in this study. It has been suggested that cognitive difficulties self-reported by patients tend to be associated with emotional difficulties and depression (49). Despite both of our SLE groups having equally high measures of depression on the CES-D, the group with higher objective impairment also reported greater subjective impairment. Because concerns regarding the use of the CES-D have been raised in this study, continued investigation using alternative measures of depression may be useful in clarifying this relationship. If future research corroborates this finding, then self-reported cognitive problems reported to the patient's rheumatologist should be considered a reflection of biologic changes, and further evaluation is highly recommended.

Consistent with prior studies, prednisone use was not associated with cognitive dysfunction in this sample of patients with SLE (7, 9, 42, 50, 51) and was unrelated to depression, pain, or fatigue. In contrast, disease activity was correlated with higher cognitive dysfunction in both SLE groups, a finding supported by some studies (52, 53), but not reported in most (7, 42, 54). This study also found that disease activity was related to self-reported pain but not to depression in patients with SLE. Results regarding disease activity related to fatigue and self-reported cognitive problems were mixed (i.e., the MAF and PAF were related to disease activity but the FSS and CFQ were not). The FSS and MAF were correlated as expected, but the MAF has more dimensions of fatigue within the scale, which may alter the overall utility of the instrument. Future studies comparing items within these measures may be useful. Similarly, more symptoms on the multidimensional PAF were associated with disease activity, but the 1-dimensional CFQ was not associated with disease activity. The CFQ has been used in a variety of studies; however, concerns regarding the ecological validity of the CFQ have been reported and may in part explain its lack of relationship with disease activity (55). Notably, the PAF total score and CFQ total score were highly correlated. Use of the PAF may be of greater value clinically because it appears to be linked to the patient's objective disease activity and describes multiple cognitive dimensions (i.e., memory, executive functions) that may be useful in treatment planning.

In this study, objective cognitive deficits were related to aspects of pain, fatigue, and depression in NPSLE and may point towards some novel treatment approaches for improved cognitive skills for these patients. For example, treating pain improved cognitive functions in a population of patients with high levels of pain, depression, and cognitive disturbance (56). Treatment approaches that have utilized psychoeducational, stress reduction, cognitive behavioral, and antidepressant therapy in patients with autoimmune disease have demonstrated success in reducing fatigue, psychological distress, and pain (57–59) and should be considered for future studies.

In conclusion, patients with SLE with a history of overt neuropsychiatric symptoms show direct relationships between cognitive impairment and depression, fatigue, pain, and subjective cognitive symptoms. This suggests multiple behavioral problems in NPSLE that may be associated. Self-reported difficulties in any of these areas should be considered important for guiding continued evaluation by the health professional. Cross-sectional and longitudinal research aimed at understanding underlying biologic mechanisms related to these behavioral changes in SLE are necessary to delineate and treat the complex nature of these behavioral disturbances.

Acknowledgements

We would like to thank all the SLE patients for their considerable time and effort contributed towards this project. We would also like to thank Sarah Forrest and Dr. Jeff Cory for their contributions in collecting the data, and Tilli Urban for her excellent administrative assistance.