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

To assess the prevalence of neurocognitive impairment (NCI) in childhood-onset systemic lupus erythematosus (cSLE) by comparing published classification criteria, and to examine associations between NCI, disease characteristics, psychosocial well-being, and intelligence.

Methods

cSLE patients and ethnicity- and age-matched healthy controls completed a neuropsychological research battery, and results were categorized by 3 different NCI classification criteria with different cutoff scores (e.g., >2, 1.5, or 1 SD below the mean) and the number of required abnormal tests or domains.

Results

Forty-one cSLE subjects and 22 controls were included. Subjects were predominantly female (∼70%) and Hispanic (∼70%). Executive functioning, psychomotor speed, and fine motor speed were most commonly affected. Method 1 classified 34.1% of cSLE subjects with NCI compared to method 2 (14.6% with decline and 7.3% with NCI) and method 3 (63.4% with NCI). The prevalence of NCI was not significantly different between the controls and patients using any of the categorization methods. NCI was not associated with SLE disease activity or characteristics or with depression. Using method 3, patients in the cognitive impairment group reported significantly lower quality of life estimates (69.7 versus 79.3; P = 0.03). Below average intellectual functioning (intelligence quotient <90) differentiated the number of test scores >1 and >1.5 SDs, but not >2 SDs below the mean.

Conclusion

NCI was prevalent in cSLE, but varied according to the chosen categorization method. A similar proportion of cSLE patients and controls had NCI, reinforcing the importance of studying an appropriate control group. Categorical classification (i.e., impaired/nonimpaired) may oversimplify the commonly observed deficits in cSLE.


INTRODUCTION

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

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease with diverse manifestations involving multiple organ systems. Neurocognitive impairment (NCI) is one of the 19 distinct neuropsychiatric syndromes of SLE defined by the American College of Rheumatology (ACR) (1), and is associated with considerable morbidity among both adults and children with SLE (1–4). Since 15–20% of all patients with SLE have the onset of disease in childhood, cognitive impairment may have a substantial impact on learning and academic and vocational success. However, a major roadblock to accurate prevalence estimates and to our understanding of neuropsychological functioning among patients with childhood-onset SLE (cSLE) can be attributed to the lack of standard criteria for NCI (2).

The ACR defines cognitive impairment as a deficit in at least 1 of the following 7 broad cognitive domains: 1) simple or complex attention, 2) memory, 3) visual–spatial processing, 4) language, 5) reasoning and problem solving, 6) psychomotor speed, and 7) executive functioning (1). Cognitive impairments may be mild to severe, represent a decline from previous functioning, and are associated with an impact on social, educational, and/or occupational functioning. The ACR definition also stipulates that neuropsychological testing is mandatory for NCI diagnosis.

Neuropsychological batteries, by definition, employ several tests to assess each of the neurocognitive domains. In research, test and/or domain aggregate scores are then used to stratify individuals into a dichotomous category of NCI (present/absent) using varying criteria as to how far scores (test and/or domain) depart from test standardization sample expectations (e.g., 1, 1.5, or 2 SDs below normative sample averages). Given this variation in NCI criteria cutoffs, it is not surprising that prior studies report a prevalence of NCI in cSLE that ranges between 20% and 95% (5–14). This broad range may also be attributed to the different number of tests and/or domains required for a diagnosis of “impairment.” Furthermore, interpretation of prevalence estimates is complicated by varying study designs and sample characteristics (i.e., patients referred for clinical assessment versus prospective research study and the inclusion of control samples).

Three recent studies with the goal of determining NCI in SLE used clearly defined but varying classification criteria methodologies (5–7). In the first method, Brunner et al employed an NCI cutoff of more than 2 SDs below the standardized mean in 1 domain, or scores between 1 and 2 SDs below the mean in 2 or more domains (5). In their study, 4 domains (memory, psychomotor speed, visual construction processing, and attention/executive functioning) were assessed using 11 tests, and 59% of the sample (research recruitment) met NCI criteria. A second method outlined (but not tested) by expert consensus methodology assigns “impairment” status to individuals who score more than 2 SDs below the normative mean in 1 or more domains and an intermediate category of “cognitive decline” to individuals achieving scores between 1.5 and 1.9 SDs below published norms in 1 or more domains (6). This publication also discussed focal impairment (impairment in 1 domain) and multifocal impairment (impairment in more than 1 domain). Most recently, a third method outlined by Muscal et al operationalized NCI as 2 or more single test scores more than 1.5 SDs below normative means spanning 2 cognitive domains (7). Seven domains (as recommended by the ACR: memory, psychomotor speed, visual construction, attention, executive functioning, intelligence, and academics) were assessed using 19 individual tests, and the prevalence of NCI ranged between 47% and 71%, depending on the cohort (prospective versus retrospective).

These studies offer formal NCI criteria for cSLE; however, their different definitions and sample characteristics as well as a lack of healthy control comparisons preclude consistent estimates of the prevalence of NCI in cSLE. This reduces the sensitivity to successfully identify NCI in cSLE and subsequently hinders appropriate intervention for children with persistent cognitive challenges. Therefore, the primary objectives of our study were to assess how the use of these different NCI criteria changes the prevalence rates in a sample of predominantly urban nonwhite cSLE patients, and to determine whether cSLE patients are significantly more likely to be categorized as having NCI than a matched control group using these differential NCI criteria methods (5–7). Our secondary objectives were to determine the associations between NCI, disease characteristics, psychosocial well-being, and intellectual level in this cohort.

Significance & Innovations

  • Although childhood-onset systemic lupus erythematosus (cSLE) patients demonstrated difficulties on individual neurocognitive measures and domains, they did not demonstrate neuropsychological deficits in excess of control participants, highlighting the importance of using an appropriate control group when examining neurocognitive impairment in demographically diverse populations.

  • It is challenging to accurately dichotomize neurocognitive impairment in cSLE, as categorical classification (i.e., impaired/nonimpaired) oversimplifies the commonly observed deficits in cSLE.

  • cSLE patients are affected particularly in the areas of executive functioning (mental flexibility), psychomotor speed, and fine motor speed, suggesting future studies focus on methodologies to overcome these difficulties.

SUBJECTS AND METHODS

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

Participants.

Individuals with cSLE followed at the Lupus Clinic at the Morgan Stanley Children's Hospital of New York-Presbyterian, Columbia University Medical Center, were eligible for recruitment as part of a prospective longitudinal study of neurocognitive function between June 2006 and August 2009. The data shown here represent the baseline neuropsychological assessment. Inclusion criteria for all participants were: 1) age 10–21 years, 2) English speaking or fluently bilingual, attending school in English for at least 2 years, and ability to complete traditional neuropsychological testing in English, and 3) absence of a comorbid condition affecting cognitive functioning (e.g., cerebral palsy, Down syndrome). Consecutive eligible patients attending the Lupus Clinic were approached until the recruitment goal of 50 subjects was attained. Five eligible patients declined participation. A control sample of age-, socioeconomic-, and ethnicity-matched individuals was recruited. Control subjects were predominantly friends of the cSLE subjects; however, a small number of healthy siblings of the cSLE subjects with no history of SLE or other autoimmune disease and healthy neighborhood controls were also recruited. The study was approved by the Institutional Research Board at Columbia University Medical Center.

Procedures.

Demographic data collected included age, sex, zip code, and family income. Screening psychosocial measures included the Beck Depression Inventory (BDI) (15) and the standard Pediatric Quality of Life Inventory (PedsQL) (16). For the cSLE subjects, disease information collected included the date of diagnosis, current medications, disease activity and damage, and any documented history of neuropsychiatric SLE as defined by the ACR case definitions (1). No subjects had previously completed a neuropsychological assessment. Disease activity was evaluated using the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI; score range 0–105) (17) and disease-specific damage scores using the Systemic Lupus International Collaborative Clinics/ACR Damage Index (SDI; score range 0–47) (18) at the rheumatology visit preceding their neuropsychological evaluation. Recent laboratory results (double-stranded DNA, complements, complete blood count, urinalysis) were collected in order to complete the disease activity measures. The presence of antiphospholipid antibodies (anticardiolipin antibodies and/or lupus anticoagulant) since cSLE diagnosis was documented.

Neuropsychological functioning was evaluated using a research battery of traditional neuropsychological tests adapted and expanded from the recommended ACR battery for assessment of adults with SLE (1) in order to be suitable for evaluation of a cSLE population. The battery assessed 7 overall cognitive domains: 1) memory, 2) language and verbal reasoning, 3) visual–spatial reasoning, 4) executive functioning, 5) psychomotor speed, 6) fine motor dexterity, and 7) academics. This is representative of the 7 domains recommended by the ACR, but differentiates fine motor control from higher-level psychomotor processes and adds assessment of academic functioning relevant to the pediatric population. Neuropsychological testing was performed by a trained psychometrist (AB) and supervised by a pediatric neuropsychologist (GSR) using multiple measures as detailed below. Assessments typically took 3–4 hours to complete. Twenty-three tests were administered to evaluate the 7 domains. Individual tests were included from the following 9 test batteries: 1) Wechsler Abbreviated Scale of Intelligence for problem solving and reasoning skills, comprised of 4 subtests (19); 2) Wide Range Assessment of Memory and Learning 2 for immediate learning and memory, comprised of 4 subtests (20); 3) letter fluency from the Delis-Kaplan Executive Function System (21); 4) Comprehensive Trail-Making Test parts 1 and 5 (22); 5) Stroop Color and Word Test, children's version (4 subtests) (23, 24); 6) Wechsler Adult Intellectual Scale III; 7) Wechsler Intellectual Scale for Children IV subtests of letter–number sequencing and digit coding (25, 26); 8) Purdue Pegboard (3 subtests) (27); and 9) Wide Range Achievement Test, Third Edition (28). In analyses examining how other neurocognitive domains varied by intellectual functioning, full-scale intellectual scores were used to compare the remaining 5 domains: 1) executive functioning, 2) memory, 3) psychomotor speed, 4) fine motor dexterity, and 5) academics. Z scores (representing SDs from the test's standardization sample mean) were generated for each test. Overall mean Z scores were then determined for each subject in each of the 7 cognitive domains. The numbers of individual tests and domain scores falling below 3 different cutoff impairment levels were also computed (e.g., >1, >1.5 and >2 SDs below the test mean).

NCI classification criteria.

NCI was operationally defined using the 3 recently published diagnostic methods, as previously outlined (5–7).

Statistical analysis.

Demographic and clinical characteristics were summarized with descriptive statistics, including the mean ± SD for continuous variables and frequencies for categorical variables. One-way analysis of variance compared patient and control groups on continuous demographic variables (i.e., age) as well as test- and domain-based standardized neurocognitive scores using a Bonferroni-corrected alpha level (P values less than 0.002 for individual test scores and P values less than 0.007 for domain scores). Pearson's chi-square for parametric data results was determined for comparisons of categorical variables (i.e., impairment, intellectual categories). As there were small numbers of subjects enrolled, we utilized parametric and nonparametric models in the analyses. All analyses were performed using the Statistical Package for the Social Sciences, version 19.

RESULTS

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

Demographics: patient sample.

Fifty cSLE patients provided informed consent and 44 completed testing. Six patients either failed to show up for multiple appointments or voluntarily withdrew consent because of a lack of time required to complete the testing. Data from 3 individuals were eliminated from the final sample due to disclosure of preexisting comorbidities (e.g., cerebral palsy, longstanding learning issues) after testing was complete. Table 1 outlines the ethnic and socioeconomic status of the participants. Subjects ranged in age from 10–21 years and the majority were female and of Hispanic ethnicity. The cSLE patients reported significantly lower quality of life as measured by the PedsQL and a trend toward more depressed symptoms than the control subjects. Mean full-scale estimated intelligent quotient (IQ) scores were within the average range for both the patient and control groups.

Table 1. Baseline demographics*
 cSLE subjects (n = 41)Control subjects (n = 22)P
  • *

    cSLE = childhood-onset systemic lupus erythematosus; ns = not significant; N/A = not applicable; BDI = Beck Depression Inventory; PedsQL = Pediatric Quality of Life Scale; WASI = Wechsler Abbreviated Intelligence Score; IQ = intelligence quotient.

  • At the time of neuropsychological testing.

Age, years   
 Mean ± SD16.4 ± 2.517.7 ± 3.00.06
 Range10.3–21.810.5–21.7 
Female sex, no. (%)30 (73)14 (64)ns
Ethnicity, no. (%)  ns
 Hispanic28 (68)16 (73) 
 African American5 (12)3 (14) 
 Asian4 (10)2 (9) 
 White3 (7)1 (5) 
 Other (mixed race/ethnicity)1 (3)1 (5) 
Yearly household income, no. (%)  ns
 <$25,00018 (44)13 (59) 
 $25,000–$49,99912 (29)4 (18) 
 $50,000–$99,9995 (12)2 (9) 
 $100,000–$150,0001 (3)0 (0) 
 >$150,0005 (12)3 (14) 
Disease duration, years   
 Mean ± SD3.6 ± 2.4N/A 
 Range0.4–12.4N/A 
Neuropsychiatric issues, no. (%)   
 Headache13 (32)N/A 
 Depression2 (5)N/A 
 Psychosis1 (3)N/A 
 Seizures1 (3)N/A 
Medications   
 Hydroxychloroquine, no. (%)40 (98)N/A 
 Prednisone, no. (%)25 (61)N/A 
 Prednisone dosage, mean (range) mg/kg/day0.23 (0.03–0.7)N/A 
 Azathioprine, no. (%)5 (12)N/A 
 Cyclophosphamide, no. (%)1 (2)N/A 
 Mycophenolate mofetil, no. (%)14 (34)N/A 
 Methotrexate, no. (%)3 (7)N/A 
BDI score (range 0–63)   
 Mean ± SD7.5 ± 6.04.4 ± 6.60.07
 Range0–230–30 
PedsQL score (range 0–100)   
 Mean ± SD73.3 ± 13.788.7 ± 9.3< 0.001
 Range46–10062–100 
Mean WASI Full-Scale IQ score   
 Mean ± SD96.8 ± 10.396.7 ± 11.1ns
 Range70–11965–116 

Comparison to normative test standardization means.

As shown in Table 2, mean standardized test scores for the cSLE subjects were lower than standardized normative values (mean ± SD Z score 0 ± 1) for 3 individual tests (Comprehensive Trail-Making Test parts 1 and 5, Purdue Pegboard subtest bilateral hand coordination) and 1 domain (psychomotor speed; P < 0.001 for each). In comparison to normative estimates of low scores, our sample had a significantly higher proportion of individuals meeting score cutoffs of more than 1 SD (56.1% versus 16%; P < 0.0001), 1.5 SDs (14.6% versus 7%; P < 0.01), and 2 SDs (7.3% versus 2%; P < 0.01) below the normative mean.

Table 2. Mean ± SD Z scores of neuropsychological tests*
Domain and testscSLE subjects, mean ± SD Z scoreControl subjects, mean ± SD Z score
  • *

    cSLE = childhood-onset systemic lupus erythematosus; WASI = Wechsler Abbreviated Scale of Intelligence; DKEFS = Delis-Kaplan Executive Function System; CTMT-5 = Comprehensive Trail-Making Test part 5; WISC = Wechsler Intelligence Scale for Children.

  • No significant differences between cSLE subjects and control subjects for all individual tests and domain scores.

  • Significant difference for comparison of mean individual test scores to published age-matched normative scores, with P value <0.001 (Bonferroni corrected for multiple [n = 23] tests).

Language and verbal reasoning  
 WASI vocabulary−0.49 ± 0.96−0.66 ± 0.86
 WASI similarities−0.12 ± 0.860.04 ± 0.74
 Overall domain score−0.21 ± 0.82−0.13 ± 0.73
Visual spatial processing  
 WASI block design−0.06 ± 0.82−0.27 ± 1.11
 WASI matrix reasoning−0.07 ± 0.730.06 ± 0.89
 Overall domain score−0.14 ± 0.51−0.10 ± 0.76
Memory  
 Story memory (immediate)−0.19 ± 0.94−0.26 ± 1.08
 List learning−0.21 ± 0.96−0.25 ± 0.79
 Design memory−0.09 ± 0.95−0.48 ± 0.93
 Picture memory−0.01 ± 0.78−0.20 ± 0.74
 Overall domain score−0.12 ± 0.64−0.30 ± 0.67
Executive functioning  
 DKEFS letter fluency−0.53 ± 1.10−0.20 ± 1.39
 CTMT-5 (part b)−0.75 ± 0.92−0.99 ± 1.11
  Stroop color word−0.10 ± 0.900.23 ± 0.82
  Stroop inhibition0.26 ± 0.700.44 ± 0.66
  WISC letter–number0.27 ± 1.040.52 ± 1.02
  Overall domain score−0.17 ± 0.550.00 ± 0.62
Psychomotor speed  
 CTMT-1−1.61 ± 1.06−1.31 ± 1.22
  WISC coding−0.05 ± 0.91−0.19 ± 0.73
  Stroop word reading−0.18 ± 0.98−0.11 ± 1.02
  Stroop color naming−0.36 ± 0.90−0.44 ± 0.80
  Overall domain score−0.55 ± 0.63−0.53 ± 0.67
Fine motor speed and dexterity  
 Purdue Pegboard, dominant hand0.05 ± 1.07−0.15 ± 0.99
 Purdue Pegboard, nondominant hand−0.38 ± 1.31−0.77 ± 1.30
 Purdue Pegboard, both hands−0.99 ± 1.11−1.19 ± 1.60
 Overall domain score−0.44 ± 1.01−0.70 ± 1.14
Academics  
 Spelling−0.04 ± 0.910.30 ± 1.16
 Math−0.43 ± 0.85−0.35 ± 1.25
 Reading0.03 ± 0.880.21 ± 1.18
 Overall domain score−0.15 ± 0.760.05 ± 1.05

Comparison to matched control sample.

There were no significant differences between cSLE patients and the control sample on any of the mean standardized individual test scores or the mean aggregate domain scores (Table 2).

Differential criteria for NCI.

As shown in Table 3, 34.1% of cSLE subjects were classified as cognitively impaired according to method 1 criteria. Using method 2, 14.6% fulfilled criteria for cognitive decline and 7.3% fulfilled criteria for cognitive impairment. Finally, using method 3, 63.4% of cSLE subjects were categorized as cognitively impaired. There were no significant differences in the proportion of impaired subjects in the cSLE and control groups using any of the 3 methods. Overall, method 3 categorized significantly more participants as being cognitively impaired than methods 1 and 2 (P < 0.0001).

Table 3. Cognitive impairment comparison by different categorization methods
 Method 1, no. (%)*Method 2a: decline, no. (%)Method 2b: impaired, no. (%)Method 3, no. (%)§
  • *

    1 domain score <2 SDs below the mean or 2 or more domain scores between 1 and 2 SDs below the mean.

  • 1 domain score <1.5 SDs below the mean.

  • 1 domain score <2 SDs below the mean.

  • §

    2 or more tests that span 2 cognitive domains with scores <1.5 SDs below the mean.

Patients14 (34.1)6 (14.6)3 (7.3)26 (63.4)
Controls8 (36.4)6 (27.3)2 (9.1)14 (63.6)

NCI classification and disease indices.

We observed no significant differences in disease duration, disease activity (SLEDAI), or damage (SDI) scores between impaired and nonimpaired groups, as shown in Table 4. However, disease activity was generally higher in the NCI group, but this difference did not reach statistical significance. The proportion of all patients with positive lupus anticoagulant (15%), anticardiolipin antibody status (42%), or renal disease (49%) also did not differ by NCI categorization methods.

Table 4. Disease indicators by neurocognitive impairment*
 Disease duration, mean ± SD yearsSLEDAI score, mean ± SDSDI score, mean ± SDLAC positive, no. (%)aCL positive, no. (%)Renal disease, no. (%)
  • *

    SLEDAI = Systemic Lupus Erythematosus Disease Activity Index; SDI = Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index; LAC = lupus anticoagulant; aCL = anticardiolipin antibody.

  • 1 domain score <2 SDs below the mean or 2 or more domain scores between 1 and 2 SDs below the mean.

  • 1 domain score <1.5 SDs below the mean.

  • §

    1 domain score <2 SDs below the mean.

  • 2 or more tests that span 2 cognitive domains with scores <1.5 SDs below the mean.

Method 1      
 No impairment4.1 ± 2.53.8 ± 3.70.4 ± 0.75 (19)10 (37)12 (44)
 Cognitive impairment3.0 ± 1.75.6 ± 5.60.6 ± 0.91 (7)7 (50)7 (50)
Method 2      
 No impairment3.9 ± 2.44.5 ± 3.80.5 ± 0.75 (16)14 (44)15 (47)
 Decline only2.6 ± 1.25.7 ± 7.50.5 ± 0.81 (17)3 (50)4 (67)
 Cognitive impairment§3.7 ± 2.60.7 ± 1.21.0 ± 1.00 (0)0 (0)0 (0)
Method 3      
 No impairment4.3 ± 2.93.6 ± 3.40.5 ± 0.74 (27)6 (40)5 (33)
 Cognitive impairment3.3 ± 1.84.9 ± 5.00.5 ± 0.82 (8)11 (42)14 (54)

NCI and psychosocial indices.

A self-administered depression index (BDI) and quality of life index (PedsQL) failed to differentiate between the NCI and unimpaired groups using method 1 or 2. Using method 3 criteria, patients in the NCI group reported significantly lower estimates of quality of life compared to the unimpaired group (Table 5). NCI was not associated with household income using any of the categorization methods (data not shown).

Table 5. Depression and quality of life by cognitive impairment*
 BDI score, mean ± SDPedsQL score, mean ± SDDepression, no. (%)
  • *

    BDI = Beck Depression Inventory; PedsQL = Pediatric Quality of Life Inventory.

  • Defined as a BDI score >10.

  • 1 domain score <2 SDs below the mean or 2 or more domain scores between 1 and 2 SDs below the mean.

  • §

    1 domain score <1.5 SDs below the mean.

  • 1 domain score <2 SDs below the mean.

  • #

    2 or more tests that span 2 cognitive domains with scores <1.5 SDs below the mean.

Method 1   
 No impairment7.4 ± 6.175.2 ± 14.06 (23)
 Cognitive impairment7.6 ± 5.869.8 ± 12.94 (29)
Method 2   
 No impairment7.5 ± 6.574.4 ± 14.28 (26)
 Decline only§5.8 ± 3.373.9 ± 12.11 (17)
 Cognitive impairment10.0 ± 4.460.8 ± 3.81 (33)
Method 3   
 No impairment6.1 ± 6.179.3 ± 11.63 (20)
 Cognitive impairment#8.2 ± 5.869.7 ± 13.87 (28)

NCI and intellectual ability.

Following previous psychometric methodology for comparing prevalence of low scores by intellectual level (29), the cSLE subject group was stratified into individuals with average range intellectual scale scores (IQ ≥90 standard score, n = 31) and those with below average scores (IQ ≤89, n = 10). As shown in Figure 1, patients with below average intelligence had significantly more test scores more than 1 and 1.5 SDs below the mean than patients with average or above intelligence (>1 SD: 7.2 versus 4.4 tests; P < 0.01 and >1.5 SDs: 3.4 versus 2.0 tests; P < 0.05). However, no differences were observed in the number of domain scores more than 1, 1.5, or 2 SDs below the mean by using this IQ cutoff (>1 SD: 0.4 versus 0.06, >1.5 SDs: 0.2 versus 0.1, >2 SDs: 0.90 versus 0.55; P = not significant for all). Similar results were observed for the control sample (data not shown).

thumbnail image

Figure 1. Test results by intelligence level. Average intelligence quotient (IQ) is defined as ≥90; below average IQ is defined as ≤89. * = P < 0.01; # = P < 0.05.

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

Using 3 different methods for categorizing NCI in cSLE, the prevalence of NCI among our sample ranged from 7.3–63.4%. Perhaps most importantly, test and domain scores as well as NCI prevalence estimates did not differ between cSLE patients and controls. Furthermore, these 3 categorization methods did not relate to measures of disease activity or damage. However, using the third method (7), NCI was associated with poorer perceived quality of life, consistent with previous research (30–34).

The criteria outlined in method 3 (2 or more single tests >1.5 SDs below the mean spanning 2 different domains) demonstrated the largest number of patients as impaired (>63%), consistent with the original authors' findings in their cohort (7). This is not surprising, given its inclusion of single test scores in the operational definition and the psychometric reality that impaired performance on 1 or more tests in test batteries with greater than 20 measures occurs in the majority of cognitively healthy subjects (35). The measurement criteria outlined in methods 1 and 2 were more moderate estimates of NCI, with each categorizing impairment in approximately one-quarter to one-third of our sample, a lower prevalence than the original study (59%) (5). Both classification methods uphold the common criterion of a domain score more than 2 SDs below the mean and also use secondary criteria to capture individuals who have less severe cognitive deficits (between 1 and 2 SDs below the mean). However, we caution researchers and clinicians in using the term “cognitive decline” (as in method 2) for studies similar to ours that assess a cross-sectional cohort and do not evaluate for an objective change in neurocognitive function over time.

Although disease activity has been linked to neurocognitive outcomes in SLE (36–40), we did not observe any association between disease duration, age at diagnosis, disease-related damage, or potential risk factors, including antiphospholipid antibodies or renal disease and NCI as assessed by any of the 3 methods, although our small sample size may have limited our power to detect these associations. Poorer quality of life was associated with the more inclusive NCI defined in method 3 and suggests the potential importance of employing greater leniency in our definitions of impairment, as well as accounting for the interrelations between mood and cognitive issues prevalent in this population (41–45).

Our sample of predominantly adolescents demonstrated neurocognitive deficits in domains previously proposed to be associated with SLE, particularly in the areas of executive functioning (mental flexibility), psychomotor speed, and fine motor speed (42, 46–50). This is consistent with proposed pathogenetic mechanisms in SLE, highlighting microstructural white matter changes, particularly in the frontal lobe, potentially explaining attentional and executive dysfunction as well as motor issues (51). Despite the salience of these deficits, cognitive interventions have rarely been evaluated in cSLE (52). In addition to educational accommodations for patients (e.g., direct instruction, modifications to environmental and task demands), specific interventions for attention problems caused by poor executive functioning are important and necessary new directions for cSLE research.

Our study also draws attention to the large number of test scores in the impaired range according to the different cutoffs, beyond those anticipated using normal curve estimates, and greater than projected base rates of abnormal scores compared to typically developing children. Although base rates of low scores are not available for this specific neuropsychological battery, one recent study demonstrated 1 or more scores >1 SD in 37.6% of the standardization sample for the Children's Memory Scale, compared to 56% of our sample meeting this impairment cutoff across all of our domains (29). Further consideration of the number of low scores found among healthy youth on the recommended pediatric cSLE test battery and the intercorrelations among these tests is strongly advised to guide future definitions of NCI within the cSLE population.

Although cSLE patients demonstrated difficulties on individual neurocognitive measures and domains, they did not demonstrate deficits in excess of control participants, nor were there any differences between the 2 groups in any of the 3 categorization systems for NCI. The similarities in low scores and NCI among our patients and healthy controls suggests overall neurocognitive difficulties in this sample, and questions if poor performance can or should be attributed to SLE. These findings highlight the importance of using an appropriate control group when examining NCI in demographically diverse populations. The use of matched control groups attempts to remedy test norms that may not reflect the diversity of socioeconomic or ethnic backgrounds among medical populations such as cSLE, a disease with greater prevalence in nonwhite populations. In the past, failure to use a matched control group has led to a disproportionate number of individuals to be classified as impaired (50, 53–56). However, some of the tests used in this study employ normative samples with increased representation of diverse cultures. For example, the Wechsler Intelligence and Delis-Kaplan Executive Function System tests have samples that include up to 25% of individuals of Hispanic origins (19, 21, 26), suggesting that our cohort (both patients and controls) may be unique, with a higher prevalence of NCI than would be expected in a control population based on available population norms. Our sample represents a predominantly urban population of lower socioeconomic status, and as such, subjects may have had fewer educational opportunities than other cSLE samples. Overall, there were few subjects in the higher household income groups, although analysis did not show fewer cases of NCI in these higher-income strata. Although recruiting friends as a comparison group controls for important sociodemographic factors, including age, ethnicity, school quality, and household income, the lack of cognitive differences among our sample may reflect how individuals seek out similarly able or challenged peers for friendships. Similarities can also be attributed to our cross-sectional methodology, as previous investigators show comparability between cSLE patients and controls early in the disease, but slower acquisition of skills and increased cognitive impairment over time (57).

To interpret neurocognitive performance in the context of the overall level of intellectual functioning, we observed that approximately 25% of cSLE patients fell in the below average intellectual group (IQ ≤89). However, below average intellect did not differentiate the number of cognitive domains scoring below 2 SDs of the mean, indicating how few subjects in both our patient and control samples met this level of impairment criteria. The number of domains and tests more than 1 and 1.5 SDs below the mean did differ by intelligence level, again suggesting more leniencies in standard impairment criteria. Although examining different test scores by dichotomizing IQ can be controversial, given the strong correlation between IQ test scores and other neurocognitive measures, it is also important to account for differential expectations of below average scores and potential patterns of impairment that may be specific to this disease process (29, 35, 58). For example, a person who is lower functioning cognitively will have lower scores and be at greater risk for misdiagnosis of NCI (i.e., false-positives) than a person who is higher functioning (and at greater risk of having a missed diagnosis, i.e., false-negatives) (29). Overall, it remains essential to assess the impact of disease on all neurocognitive domains, including intellectual indices, and to acknowledge the additional challenges inherent in assessing NCI in cSLE patients whose cognitive skills are not yet fully developed.

Several limitations of the present study should be noted. Most saliently, the small sample size limited the statistical power of our analyses. Our predominantly Hispanic sample with limited inclusion of other sociodemographic groups also restricts generalizability to other more diverse cSLE populations. As with many neurocognitive cSLE studies, this battery of neuropsychological tests was different than previous studies, yielding a different number of single test scores and reflective of slightly different domains. After the initiation of this study, a recommended test battery for cSLE was published addressing the necessity for pediatric-specific instruments (4). Although our study used many of the same as well as comparable tests, we recommend that future research follow published guidelines as a minimum to permit increased comparability across studies. Finally, our study does not elaborate on the causality of NCI beyond our examination of disease characteristics, or examine any potential neuroanatomic differences among our sample.

Despite these limitations, by comparing existing criteria for NCI among cSLE patients in the same sample, our study controls for previous variability in differential neuropsychological measures and methodologic differences in recruitment in assessing NCI. Our results highlight the challenges of accurately dichotomizing NCI generally and more specifically in a disease such as cSLE known for its diversity in presentation and impact. Further consideration and refinement of the concept of NCI among children with SLE is needed. Future efforts are encouraged to go beyond dichotomous ranking systems and to continue to consider intelligence as well as other functional outcomes. Overall, consistent use of a single definition in research studies using similar test batteries based on these considerations will provide further data for assessment of impairment criteria. Our results suggest the importance of using domain-based scores and flexibility in cutoff criteria between 1 and 1.5 SDs below the mean, given how few individuals met thresholds of >2 SDs. Further, addressing the cumulative number of impaired domains and addressing which neurocognitive processes these represent may be most valuable in elucidating the underlying neuropsychological networks associated with cSLE. We are currently testing this in a new multiethnic sample to further consider and expand upon these issues, as well as to consider potential cognitive intervention in this population. Importantly, no definition should be used in an “all or nothing” manner for the decision of when to pursue educational assistance and closer followup. We remain optimistic that future multicenter studies using common methodology and assessment batteries will be able to accomplish these goals and provide valuable insight into neuropsychological development and quality of life among children and adolescents with 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. Levy 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. Williams, Aranow, Ross, Imundo, Kahn, Diamond, Levy.

Acquisition of data. Barsdorf, Imundo, Eichenfield, Kahn, Levy.

Analysis and interpretation of data. Williams, Aranow, Ross, Barsdorf, Levy.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES
  • 1
    ACR Ad Hoc Committee on Neuropsychiatric Lupus Nomenclature. The American College of Rheumatology nomenclature and case definitions for neuropsychiatric lupus syndromes. Arthritis Rheum 1999; 42: 599608.
  • 2
    Benseler SM, Silverman ED. Neuropsychiatric involvement in pediatric systemic lupus erythematosus [review]. Lupus 2007; 16: 56471.
  • 3
    Kozora E, Hanly JG, Lapteva L, Filley CM. Cognitive dysfunction in systemic lupus erythematosus: past, present, and future [review]. Arthritis Rheum 2008; 58: 328698.
  • 4
    Levy DM, Ardoin SP, Schanberg LE. Neurocognitive impairment in children and adolescents with systemic lupus erythematosus. Nat Clin Pract Rheumatol 2009; 5: 10614.
  • 5
    Brunner HI, Ruth NM, German A, Nelson S, Passo MH, Roebuck-Spencer T, et al. Initial validation of the pediatric automated neuropsychological assessment metrics for childhood-onset systemic lupus erythematosus. Arthritis Rheum 2007; 57: 117482.
  • 6
    Mikdashi J, Esdaile J, Alarcon G, Crofford L, Fessler BJ, Shanberg L, et al. Proposed response criteria for neurocognitive impairment in systemic lupus erythematosus clinical trials. Lupus 2007; 16: 41825.
  • 7
    Muscal E, Bloom D, Hunter J, Myones B. Neurocognitive deficits and neuroimaging abnormalities are prevalent in children with lupus: clinical and research experiences at a US pediatric institution. Lupus 2009; 19: 26879.
  • 8
    Levy DM, Ross GS. Cognitive and psychiatric manifestations of childhood-onset systemic lupus erythematosus. In: Nass R, Frank Y, editors. Cognitive and behavioral manifestations of pediatric diseases. New York: Oxford University Press; 2010. p. 3046.
  • 9
    Sibbitt WL Jr, Brandt JR, Johnson CR, Maldonado ME, Patel SR, Ford CC, et al. The incidence and prevalence of neuropsychiatric syndromes in pediatric onset systemic lupus erythematosus. J Rheumatol 2002; 29: 153642.
  • 10
    Harel L, Sandborg C, Lee T, von Scheven E. Neuropsychiatric manifestations in pediatric systemic lupus erythematosus and association with antiphospholipid antibodies. J Rheumatol 2006; 33: 18737.
  • 11
    Yancey CL, Doughty RA, Athreya BH. Central nervous system involvement in childhood systemic lupus erythematosus. Arthritis Rheum 1981; 24: 138995.
  • 12
    Steinlin MI, Blaser SI, Gilday DL, Eddy AA, Logan WJ, Laxer RM, et al. Neurologic manifestations of pediatric systemic lupus erythematosus. Pediatr Neurol 1995; 13: 1917.
  • 13
    Loh WF, Hussain IM, Soffiah A, Lim YN. Neurological manifestations of children with systemic lupus erythematosus. Med J Malaysia 2000; 55: 45963.
  • 14
    Olfat MO, Al-Mayouf SM, Muzaffer MA. Pattern of neuropsychiatric manifestations and outcome in juvenile systemic lupus erythematosus. Clin Rheumatol 2004; 23: 3959.
  • 15
    Beck AT, Steer RA. Manual for the Beck Depression Inventory. San Antonio: Psychological Corporation; 1987.
  • 16
    Varni J, Seid M, Rode C. The PedsQL measurement model for the pediatric quality of life inventory. Med Care 1999; 37: 12639.
  • 17
    Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang DH, and the Committee on Prognosis Studies in SLE. Derivation of the SLEDAI: a disease activity index for lupus patients. Arthritis Rheum 1992; 35: 63040.
  • 18
    Gladman D, Ginzler E, Goldsmith C, Fortin P, Liang M, Urowitz M, et al. The development and initial validation of the Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index for systemic lupus erythematosus. Arthritis Rheum 1996; 39: 3639.
  • 19
    Wechsler D. Wechsler Abbreviated Scale of Intelligence (WASI). San Antonio: Harcourt Assessment; 1999.
  • 20
    Adams W, Sheslow D. Wide range assessment of memory and learning (WRAML). 2nd ed. Wilmington: Wide Range; 2003.
  • 21
    Delis D, Kaplan E, Kramer J. Delis-Kaplan Executive Function System. San Antonio: Pearson, PsychCorp; 2001.
  • 22
    Reynolds C. Comprehensive Trail-Making Test. Austin: Pro-Ed; 2002.
  • 23
    Golden C, Freshwater S. Stroop Color and Word Test. Wood Dale: Stoelting; 2002.
  • 24
    Golden C. Stroop Color and Word Test children's version for ages 5-14. Wood Dale: Stoelting; 2004.
  • 25
    The Psychological Corporation. WAIS-WMS-II technical manual. San Antonio: Author; 1997.
  • 26
    Wechsler D. Wechsler Intelligence Scale for Children (WISC). 4th ed. San Antonio: Harcourt Assessment; 2004.
  • 27
    Tiffin J, Asher E. The Purdue Pegboard: norms and studies of reliability and validity. J Appl Psychol 1948; 32: 23447.
  • 28
    Wilkinson G. Wide Range Achievement Test 3. Wilmington: Wide Range; 1993.
  • 29
    Brooks BL, Iverson GL, Sherman EM, Holdnack JA. Healthy children and adolescents obtain some low scores across a battery of memory tests. J Int Neuropsychol Soc 2009; 15: 6137.
  • 30
    McElhone K, Abbott J, Teh LS. A review of health related quality of life in systemic lupus erythematosus. Lupus 2006; 15: 63343.
  • 31
    Moorthy LN, Harrison MJ, Onel KB, Lehman TJA. Relationship of quality of life and physical function measures with disease activity in children with systemic lupus erythematosus. Lupus 2005; 14: 2807.
  • 32
    Moorthy LN, Peterson MG, Hassett AL, Baratelli M, Chalom EC, Hashkes PJ, et al. Relationship between health-related quality of life and SLE activity and damage in children over time. Lupus 2009; 18: 6229.
  • 33
    Dobkin PL, Da Costa D, Drista M, Fortin PR, Senecal JL, Goulet JR, et al. Quality of life in systemic lupus erythematosus patients during more and less active disease states: differential contributors to mental and physical health. Arthritis Care Res 1999; 12: 40110.
  • 34
    Brunner HI, Higgins GC, Wiers K, Lapidus SK, Olson JC, Onel K, et al. Health-related quality of life and its relationship to patient disease course in childhood-onset systemic lupus erythematosus. J Rheumatol 2009; 36: 153645.
  • 35
    Binder LM, Iverson GL, Brooks BL. To err is human: “abnormal” neuropsychological scores and variability are common in healthy adults. Arch Clin Neuropsychol 2009; 24: 3146.
  • 36
    Denburg SD, Carbotte RM, Ginsberg JS, Denburg JA. The relationship of antiphospholipid antibodies to cognitive function in patients with systemic lupus erythematosus. J Int Neuropsychol Soc 1997; 3: 37786.
  • 37
    Afeltra A, Garzia P, Mitterhofer AP, Vadacca M, Galluzzo S, Del Porto F, et al. Neuropsychiatric lupus syndromes: relationship with antiphospholipid antibodies. Neurology 2003; 61: 10810.
  • 38
    Alarcon-Segovia D, Estanol B, Garcia-Ramos G, Villa AR. Antiphospholipid antibodies and the antiphospholipid syndrome: clinical relevance in neuropsychiatric systemic lupus erythematosus. Ann N Y Acad Sci 1997; 823: 27988.
  • 39
    Emori A, Matsushima E, Aihara O, Ohta K, Koike R, Miyasaka N, et al. Cognitive dysfunction in systemic lupus erythematosus. Psychiatry Clin Neurosci 2005; 59: 5849.
  • 40
    Denburg SD, Denburg JA. Cognitive dysfunction and antiphospholipid antibodies in systemic lupus erythematosus. Lupus 2003; 12: 88390.
  • 41
    Lapteva L, Nowak M, Yarboro CH, Takada K, Roebuck-Spencer T, Weickert T, et al. Anti–N-methyl-D-aspartate receptor antibodies, cognitive dysfunction, and depression in systemic lupus erythematosus. Arthritis Rheum 2006; 54: 250514.
  • 42
    Monastero R, Bettini P, Del Zotto E, Cottini E, Tincani A, Balestrieri G, et al. Prevalence and pattern of cognitive impairment in systemic lupus erythematosus patients with and without overt neuropsychiatric manifestations. J Neurol Sci 2001; 184: 339.
  • 43
    Denburg SD, Carbotte RM, Denburg JA. Cognitive impairment in systemic lupus erythematosus: a neuropsychological study of individual and group deficits. J Clin Exp Neuropsychol 1987; 9: 32339.
  • 44
    Sonies B, Klippel J, Gerber R, Gerber C. Cognitive performance in systemic lupus erythematosus [abstract]. Arthritis Rheum 1982; 25 Suppl: S80.
  • 45
    Kozora E, Ellison MC, West S. Depression, fatigue, and pain in systemic lupus erythematosus (SLE): relationship to the American College of Rheumatology SLE neuropsychological battery. Arthritis Rheum 2006; 55: 62835.
  • 46
    Luna B, Garver KE, Urban TA, Lazar NA, Sweeney JA. Maturation of cognitive processes from late childhood to adulthood. Child Dev 2004; 75: 135772.
  • 47
    DiFrancesco MW, Holland SK, Ris MD, Adler CM, Nelson S, DelBello MP, et al. Functional magnetic resonance imaging assessment of cognitive function in childhood-onset systemic lupus erythematosus: a pilot study. Arthritis Rheum 2007; 56: 415163.
  • 48
    Hanly JG, Walsh NM, Sangalang V. Brain pathology in systemic lupus erythematosus. J Rheumatol 1992; 19: 73241.
  • 49
    Fisk JD, Eastwood B, Sherwood G, Hanly JG. Patterns of cognitive impairment in patients with systemic lupus erythematosus. Br J Rheumatol 1993; 32: 45862.
  • 50
    American Educational Research Association, American Psychological Association, National Council on Measurement in Education. Standards for educational and psychological testing. New York: American Psychological Association; 1999.
  • 51
    Filley CM. White matter: organization and functional relevance. Neuropsychol Rev 2010; 20: 15873.
  • 52
    Harrison MJ, Morris KA, Horton R, Toglia J, Barsky J, Chait S, et al. Results of intervention for lupus patients with self-perceived cognitive difficulties. Neurology 2005; 65: 13257.
  • 53
    Ardila A, Rosselli M, Puente M. Neuropsychological evaluation of the Spanish speaker. New York: Plenum Press; 1994.
  • 54
    Ferraro F. Minority and cross-cultural aspects of neuropsychological assessments. Herreweg, Lisse (The Netherlands): Swets & Zeitlinger; 2001.
  • 55
    Figueroa R. Psychological testing of linguistic-minority students: knowledge gaps and regulations. Except Child 1989; 56: 14552.
  • 56
    Duran R. Assessment and instruction of at-risk Hispanic students. Except Child 1989; 56: 1548.
  • 57
    Klein-Gitelman MS, Zelko F, Kress A, Hunter S, Wagner-Weiner L. Comparison of neuro-cognitive function in children with pediatric systemic lupus erythematosus (pSLE) and their peers: second year follow up [abstract]. Arthritis Rheum 2002; 46 Suppl: S216.
  • 58
    Brooks BL, Strauss E, Sherman EM, Iverson GL, Slick DJ. Developments in neuropsychological assessment: refining psychometric and clinical interpretive methods. Can Psychol 2009; 50: 196209.