SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Objective

Cognitive dysfunction and cardiovascular disease are common and debilitating manifestations of systemic lupus erythematosus (SLE). In this study, we evaluated the relationship between cardiovascular events, traditional cardiovascular risk factors, and SLE-specific risk factors as predictors of cognitive dysfunction in a large cohort of participants with SLE.

Methods

Subjects included 694 participants from the Lupus Outcomes Study (LOS), a longitudinal study of SLE outcomes based on an annual telephone survey querying demographic and clinical variables. The Hopkins Verbal Learning Test-Revised and the Controlled Oral Word Association Test were administered to assess cognitive function. Multiple logistic regression was used to identify cardiovascular events (myocardial infarction, stroke), traditional cardiovascular risk factors (hypertension, hyperlipidemia, diabetes mellitus, obesity, smoking), and SLE-specific risk factors (antiphospholipid antibodies [aPL], disease activity, disease duration) associated with cognitive impairment in year 7 of the LOS.

Results

The prevalence of cognitive impairment as measured by verbal memory and verbal fluency metrics was 15%. In adjusted multiple logistic regression analyses, aPL (odds ratio [OR] 2.10, 95% confidence interval [95% CI] 1.3–3.41), hypertension (OR 2.06, 95% CI 1.19–3.56), and a history of stroke (OR 2.27, 95% CI 1.16–4.43) were significantly associated with cognitive dysfunction. In additional analyses evaluating the association between these predictors and severity of cognitive impairment, stroke was significantly more prevalent in participants with severe impairment when compared to those with mild or moderate impairment (P = 0.036).

Conclusion

These results suggest that the presence of aPL, hypertension, and stroke are key variables associated with cognitive impairment, which may aid in identification of patients at greatest risk.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Cognitive dysfunction is a common neuropsychiatric manifestation of systemic lupus erythematosus (SLE), with a prevalence ranging from 21–81% (1, 2). While existing studies have identified a combination of biologic and socioeconomic factors as potential predictors of cognitive impairment, the etiology of cognitive impairment in SLE remains unclear (3, 4). There is substantial evidence linking cardiovascular disease and risk factors identified in the Framingham Heart Study with cognitive dysfunction in the general population. In the Whitehall II Study, for example, coronary heart disease was associated with lower cognitive performance in middle-aged individuals (5). Another study correlated hypertension and diabetes mellitus with cognitive decline in middle-aged adults (6). Similar to the general population, prior SLE studies have also implicated hypertension and diabetes mellitus as potential predictors of cognitive dysfunction (3, 4). However, despite the increased burden of premature cardiovascular disease in patients with SLE, the relationship between cardiovascular risk factors and events and cognitive dysfunction has not been fully explored (7–9).

The purpose of this study was to investigate the relationships between cardiovascular events (myocardial infarction [MI], stroke), traditional cardiovascular risk factors (hypertension, hyperlipidemia, diabetes mellitus, obesity, smoking), SLE-specific risk factors (antiphospholipid antibodies [aPL], disease activity, disease duration), and cognitive dysfunction in a large cohort of well-characterized individuals with SLE. We aimed to identify the relationship of specific cardiovascular risk factors or events to cognitive dysfunction beyond sociodemographic and disease characteristics.

Significance & Innovations

  • There is in an increased burden of premature cardiovascular disease as well as a high prevalence of cognitive dysfunction in patients with systemic lupus erythematosus (SLE).

  • Despite substantial evidence linking cardiovascular disease with cognitive dysfunction in the general population, the relationship has not been fully explored in patients with SLE.

  • Antiphospholipid antibodies, hypertension, and a history of stroke were independently associated with cognitive dysfunction.

  • Stroke was significantly more prevalent in participants with severe cognitive impairment.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Subjects

The Lupus Outcomes Study (LOS) is a large cohort of individuals with SLE followed longitudinally through a structured annual telephone interview conducted by trained survey workers. The year-to-year retention rate in the LOS is 92%. Diagnoses of all participants enrolled in the study were confirmed by medical chart review (10). Details regarding subject recruitment have been previously described (11). Briefly, initial subject recruitment occurred between 2002 and 2004, with a second enrollment period beginning in 2006. The participants were recruited through academic medical centers (25%), community rheumatology practices (11%), and various community-based sources, including support groups, conferences, and other media (64%). Of 957 total participants enrolled at the time, 694 were included in this study because they had a complete data set, including cognitive function outcomes in year 7, demographics, medical history, depression, disease activity, and at least 1 available aPL test. The research protocol has been approved by the University of California, San Francisco Committee on Human Research and all participants gave their informed consent prior to participation.

Data

Data were derived from 2 sources. The first was the annual structured telephone interviews, containing validated measures pertaining to demographic and socioeconomic characteristics, SLE disease activity and manifestations, medications, general health and comorbidities, depression, employment, health care utilization, and health insurance coverage. The second was medical record reviews to obtain disease duration and laboratory test results.

Measures.

Cognitive function.

LOS participants were screened for memory impairment by telephone interview using the Hopkins Verbal Learning Test-Revised (HVLT-R), a valid and reliable measure of verbal learning and memory (12, 13). The test consists of a 12-item word list that is presented on 3 successive learning trials and a delayed recall trial, yielding a total recall score and a delayed recall score. The Controlled Oral Word Association Test was employed as a measure of verbal fluency (14, 15). This test consists of 3 trials for which participants generate words beginning with specific letters under timed conditions (15). These measures have been validated recently against a comprehensive cognitive battery and were found to be of sufficient reliability for the detection of cognitive impairment in SLE (16). Age-stratified Z scores were derived for HVLT recall, HVLT delayed recall, and verbal fluency tests. For the purpose of classifying participants and assessing overall cognitive function, we employed a commonly utilized method of data reduction and created a composite Z score index in which Z scores from all 3 testing domains were averaged (17–19). Based on their composite Z score, participants were considered to have intact cognitive function if they scored better than −1.0 SD below population norms, and were considered to have cognitive impairment if they scored worse than or equal to −1.0 SD below age-stratified normative values.

Cardiovascular risk, events, and disease-related variables

Cardiovascular events included self-reported MI or stroke in years 1–7 of the LOS. Cardiovascular risk factors included self-reported hypertension, hyperlipidemia, and diabetes mellitus in each year. Obesity was defined by current body mass index ≥30 kg/m2, and smoking status was defined as current smoking.

Disease-related factors included disease activity, disease duration, and aPL. Based on medical record reviews, participants were considered to be positive for aPL if they had at least 1 test result indicating the presence of anticardiolipin antibodies (IgG or IgM) or anti–β2-glycoprotein I antibodies (IgG or IgM), or a lupus anticoagulant measured by an abnormal Russell's viper venom test result at least one point in time. SLE-related disease activity was measured by the Systemic Lupus Activity Questionnaire (SLAQ) score, a validated measure of disease activity in SLE accounting for constitutional, mucocutaneous, and musculoskeletal symptoms and other domains (20, 21). Disease duration was determined from the date of SLE diagnosis confirmed by medical records.

Covariates

Sociodemographic factors included age, sex, education level (high school or less, some college, college graduate, and postgraduate), and living below poverty thresholds (household income below 125% of the Federal Poverty Level). The Center for Epidemiologic Studies Depression Scale (CES-D), a measure of depressive symptom severity, was used to evaluate depression status. Depression was defined as a CES-D score ≥24, as previously described (22, 23).

Statistical analysis

Bivariate comparisons were conducted between impaired and unimpaired individuals on sociodemographic factors, disease activity and duration, depression, aPL, and cardiovascular risk factors and events. Chi-square tests and analysis of variance were used as appropriate to identify differences between the groups. We used unadjusted and hierarchical adjusted logistic regression to identify cardiovascular risk factors and events as potential predictors of cognitive dysfunction. The adjusted analyses controlled for sociodemographics (we did not control for age because the cognitive function measures were already age adjusted) and depression. Variables of interest were evaluated in a hierarchical fashion, with disease-related variables (aPL, disease activity, disease duration), cardiovascular risk factors (hypertension, hyperlipidemia, diabetes mellitus, obesity, current smoking), and cardiovascular events (MI, stroke) added successively.

A sensitivity analysis was conducted, including a modified SLAQ index that excluded symptoms that could overlap with cognitive dysfunction (e.g., concentration problems, depression, reduced energy), and this model yielded highly comparable results (data not shown); therefore, the original measure was included in the final model to maintain the psychometric integrity of the SLAQ. Collinearity thresholds were also considered. In additional analyses, we divided the cognitively impaired group into mild- moderate impairment (Z score less than or equal to −1.0 but greater than −2.0) and severe impairment (Z score less than or equal to −2.0). We used chi-square tests to identify differences between the groups for variables identified as significant in the previous adjusted logistic regression.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Six hundred ninety-four participants were included in this study. Based on the composite cognitive function score, 107 (15%) were classified as having cognitive impairment, whereas 587 (85%) were considered to be unimpaired. Subject characteristics are shown in Table 1. Participants with cognitive impairment were more likely to have a lower level of educational attainment and to be living below the poverty thresholds. Participants with cognitive dysfunction were also more likely to have CES-D scores indicative of depression. Disease duration was similar in both of the groups, but participants with cognitive impairment had a significantly greater SLAQ index, indicative of greater disease activity. With regard to cardiovascular risk factors and events, participants with cognitive dysfunction were more likely to report diabetes mellitus, hypertension, and a history of stroke. They were also more likely to be a current smoker and to have laboratory evidence of aPL positivity. There were no significant differences between the groups with regard to MI, hyperlipidemia, and obesity.

Table 1. Subject characteristics stratified by cognitive function*
 All patients (n = 694)Cognitive function
Unimpaired (n = 587)Impaired (n = 107)P
  • *

    Values are the number (percentage) unless otherwise indicated. All variables are self-reported with the exception of presence of antiphospholipid antibodies and disease duration, which were collected from chart review. SLAQ = Systemic Lupus Activity Questionnaire.

  • Significant.

Age, mean ± SD years50.0 ± 12.950.3 ± 12.648.4 ± 14.50.061
Women640 (92.2)542 (92.3)98 (91.6)0.791
Education    
 High school or less114 (16.4)77 (13.1)37 (34.6)0.000
 Some college282 (40.6)230 (39.2)52 (48.6) 
 College graduate173 (24.9)160 (27.3)13 (12.1) 
 Postgraduate (ref.)125 (18)120 (20.4)5 (4.7) 
Below poverty96 (13.8)64 (10.9)32 (29.9)0.000
Depression129 (18.6)79 (13.5)50 (46.7)0.000
Disease activity (SLAQ), mean ± SD11.4 ± 7.810.6 ± 7.415.6 ± 8.70.025
Disease duration, mean ± SD years17.1 ± 8.617.4 ± 8.615.7 ± 8.50.866
Antiphospholipid antibodies274 (39.5)217 (37)57 (53.3)0.002
Hypertension401 (57.8)325 (55.4)76 (71)0.003
Hyperlipidemia276 (39.8)233 (39.7)43 (40.2)0.924
Diabetes mellitus77 (11.1)59 (10.1)18 (16.8)0.040
Obese189 (27.2)153 (26.1)36 (33.6)0.105
Current smoking61 (8.8)43 (7.3)18 (16.8)0.001
Myocardial infarction46 (6.6)37 (6.3)9 (8.4)0.420
Stroke71 (10.2)50 (8.5)21 (19.6)0.000

Unadjusted and hierarchical adjusted logistic regression analyses predicting cognitive dysfunction are shown in Table 2. In the unadjusted model, hypertension (odds ratio [OR] 1.98, 95% confidence interval [95% CI] 1.26–3.09), diabetes mellitus (OR 1.81, 95% CI 1.02–3.21), current smoking (OR 2.56, 95% CI 1.41–4.63), and stroke (OR 2.62, 95% CI 1.50–4.58) were all associated with cognitive impairment. Disease activity (OR 1.08, 95% CI 1.05–1.11) and aPL (OR 1.94, 95% CI 1.28–2.94) were also associated with cognitive impairment in the unadjusted model. Disease duration, hyperlipidemia, obesity, and history of MI were not significantly associated with cognitive impairment.

Table 2. Hierarchical logistic regression evaluating cardiovascular risk factors and events as predictors of cognitive dysfunction (n = 694 observations)*
 Unadjusted model, OR (95% CI)Adjusted models
Model 1, OR (95% CI)Model 2, OR (95% CI)§Model 3, OR (95% CI)
  • *

    All variables are self-reported with the exception of presence of antiphospholipid antibodies and disease duration, which were collected from chart review. OR = odds ratio; 95% CI = 95% confidence interval; SLAQ = Systemic Lupus Activity Questionnaire; aPL = antiphospholipid antibodies; MI = myocardial infarction.

  • All models adjusted for sex, education, poverty status, and depression.

  • Disease-related variables.

  • §

    Disease-related variables and cardiovascular risk factors.

  • Disease-related variables, cardiovascular risk factors, and events.

  • #

    Significant to P < 0.05.

Disease activity (SLAQ)1.08 (1.05–1.11)#1.04 (1.01–1.07)#1.04 (1.00–1.07)1.04 (1.00–1.07)
Disease duration0.98 (0.95–1.00)0.98 (0.95–1.01)0.97 (0.94–1.00)0.97 (0.94–1.00)
aPL1.94 (1.28–2.94)#2.14 (1.34–3.41)#2.24 (1.39–3.61)#2.10 (1.3–3.41)#
Hypertension1.98 (1.26–3.09)# 2.08 (1.21–3.58)#2.06 (1.19–3.56)#
Hyperlipidemia1.02 (0.67–1.55) 0.81 (0.49–1.35)0.74 (0.43–1.26)
Diabetes mellitus1.81 (1.02–3.21)# 1.35 (0.66–2.76)1.24 (0.6–2.57)
Obese1.44 (0.93–2.24) 0.68 (0.39–1.17)0.71 (0.41–1.24)
Current smoking2.56 (1.41–4.63)# 1.47 (0.73–2.95)1.51 (0.74–3.07)
MI1.37 (0.64–2.92)  0.96 (0.38–2.44)
Stroke2.62 (1.50–4.58)#  2.27 (1.16–4.43)#

In hierarchical logistic regression adjusted for sex, education, poverty status, and depression, 3 models evaluated potential predictors of cognitive dysfunction. Model 1 included disease-related variables, model 2 added general cardiovascular risk factors, and model 3 added a history of cardiovascular events. In the final model including all potential predictors, aPL (OR 2.10, 95% CI 1.3–3.41), hypertension (OR 2.06, 95% CI 1.19–3.56), and a history of stroke (OR 2.27, 95% CI 1.16–4.43) remained significantly associated with cognitive dysfunction. Disease activity, diabetes mellitus, and current smoking were no longer significantly associated with cognitive impairment.

In order to explore the association between significant predictors and severity of cognitive impairment, additional analyses were performed. Characteristics of participants with mild-moderate and severe cognitive impairment are shown in Table 3. While aPL and hypertension were associated with overall cognitive impairment in the previous analysis, subset analysis did not show significant differences between varying levels of cognitive impairment. Stroke, however, was significantly more prevalent in severely impaired participants when compared with mildly-moderately impaired participants (P = 0.036). Results are graphically shown in Figure 1, where it can be noted that the relative prevalence of stroke is substantially increased.

Table 3. Prevalence of risk factors and events in patients with mild-moderate and severe cognitive impairment*
 Mild-moderate cognitive impairment (n = 76)Severe cognitive impairment (n = 31)P
  • *

    Values are the number (percentage). The mild-moderate cognitive impairment group represents Z scores less than or equal to −1.0 but greater than −2.0, whereas the severe cognitive impairment group represents Z scores less than or equal to −2.0. aPL = antiphospholipid antibodies.

  • Significant.

aPL40 (52.6)17 (54.8)0.836
Hypertension51 (67.1)25 (80.6)0.161
Stroke11 (14.5)10 (32.3)0.036
thumbnail image

Figure 1. Prevalence of cardiovascular risk factors and events in patients with mild-moderate and severe cognitive impairment relative to unimpaired patients.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

In this study, we sought to identify cardiovascular risk factors and events that are potentially associated with cognitive dysfunction in SLE. We found that hypertension, stroke, and the presence of aPL were significantly associated with cognitive impairment. In addition, we found that stroke was significantly more prevalent in participants with severe dysfunction when compared to participants with mild-moderate dysfunction.

The relationship between hypertension and cognitive dysfunction in our study was robust and is consistent with at least 1 prior study in SLE (3). Moreover, hypertension has been well established as a predictor of cognitive dysfunction in the general cardiovascular literature. In a large cohort study of more than 10,000 patients, essential hypertension at baseline was associated with cognitive decline over the following 6 years (6). Furthermore, similar to the cognitive dysfunction seen in our study, the Veterans Affairs Normative Aging Study demonstrated that individuals with uncontrolled hypertension exhibited decrements in recall and verbal fluency with increasing age when compared with normotensive patients (24). Although the mechanism linking hypertension with cognitive dysfunction in patients with or without SLE is not entirely understood, studies have correlated hypertension with increased white matter hyperintensities and brain atrophy, specifically of the prefrontal cortex (25, 26). Although hypertension has been previously reported as a predictor of ischemic stroke in patients with SLE, our data suggest an association that may be in part independent of reported stroke (27, 28).

aPL positivity was also significantly associated with cognitive impairment in our adjusted logistic regression. Although aPL previously have been found to be associated with cognitive impairment, it remains unclear whether aPL contribute to cognitive decline through a direct pathogenic effect on neurons or through thrombotic or other mechanisms (3, 4, 28–30). aPL have a well-established association with stroke (28). Stroke is also a well-known cause of cognitive dysfunction in the general population (31, 32). Although an infrequent occurrence, stroke was significantly associated with cognitive dysfunction in our study. Among individuals with documented persistently positive aPL in our cohort, there was also an insignificant trend toward an increased frequency of self-reported stroke, which suggests that they may contribute to cognitive dysfunction (data not shown). The association with severe impairment identified in our study suggests that stroke may represent an end point on a spectrum of disease pathology that leads to cognitive dysfunction, at which point outcomes are more severe. Given the interrelatedness of stroke with aPL and hypertension, early identification of the latter two may enable clinicians to identify patients at greatest risk for cognitive decline and therapeutically intervene.

It is important to note several limitations in our study when evaluating the contribution of aPL. Although our study was designed to evaluate the relationship between Framingham-type cardiovascular risk factors and cognitive dysfunction using self-report data, we controlled for aPL using a single positive test. In doing so, we are unable to fully assess the contribution of specific aPL, titers of antibodies, multiple positive antibodies, or true aPL syndrome (APS) in our cohort. APS is most widely defined using the revised Sapporo criteria, which include 2 positive antibody tests documented at least 12 weeks apart plus the presence of documented vascular thrombosis or pregnancy morbidity (33). In subset analyses evaluating patients who had data from 2 successive tests and documented clinical evidence of thrombosis or pregnancy morbidity, we were only able to identify 15 patients who met strict criteria for APS, a cohort that was underpowered to be evaluated with respect to cognitive function (data not shown).

Another major limitation of our study is the fact that cardiovascular risk factors and events were evaluated via self-report. Although self-report has been deemed a valid and reliable method of data acquisition for some variables (e.g., hypertension), it has been suggested that the validity of self-report for other variables (e.g., stroke) is low and influenced by cognitive function (34, 35). It is certainly possible that self-report was significantly more limited in the patients with cognitive dysfunction, leading to underreporting of risk factors and perhaps biasing the results of our study toward the null hypothesis. Furthermore, severity, timing, and therapeutic interventions were not queried in association with cardiovascular risk factors and events. Therefore, we do not know the temporal relationship between cognitive dysfunction and stroke, which could be critical to understanding the pathogenesis. There are also limitations to our cognitive dysfunction outcome variable. Although these tests capture several “subcortical” cognitive domains observed to be impaired in lupus, they certainly do not capture all possible cognitive deficits (1, 19, 36, 37). As a result, our prevalence rates of cognitive dysfunction were probably reduced in comparison to studies investigating the full breadth of cognitive domains, again potentially biasing our results toward the null. Nevertheless, this is the first study to comprehensively evaluate cardiovascular risk factors and events in association with cognitive dysfunction in a large cohort of patients with SLE.

Cognitive dysfunction is a major cause of morbidity in patients with SLE, but the etiology has not been well established. Cardiovascular disease has emerged as a significant predictor of the development of other aspects of neuropsychiatric SLE, such as depression (38). An understanding of the relationship between cardiovascular risk factors and cognitive dysfunction may allow us to identify patients at greater risk for decline. Importantly, many of these potential risk factors for cognitive dysfunction, including hypertension, are potentially modifiable. Future longitudinal studies are necessary to clarify the causality of these relationships. A better understanding of specific targets for prevention and intervention can serve to minimize the burden of one of the most common neuropsychiatric syndromes occurring in SLE.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Murray 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 conception and design. Murray, Yazdany, Yelin, Julian.

Acquisition of data. Kaiser, Criswell, Trupin, Yelin, Katz, Julian.

Analysis and interpretation of data. Murray, Yazdany, Kaiser, Criswell, Trupin, Yelin, Julian.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES