a longitudinal study of cognitive functioning in schizophrenia: clinical and biological predictors

Schizophrenia Research xxx (2014) xxx–xxx

SCHRES-05842; No of Pages 6

Contents lists available at ScienceDirect

Schizophrenia Research

j ourna l homepage: www.e lsev ie r .com/ locate /schres

A longitudinal study of cognitive functioning in schizophrenia: Clinicaland biological predictors

Faith Dickerson a,⁎, Jennifer Schroeder b, Cassie Stallings a, Andrea Origoni a, Emily Katsafanas a,Lucy A.B. Schwienfurth a, Christina L.G. Savage a, Sunil Khushalani a, Robert Yolken c

a Stanley Research Program, Sheppard Pratt Health System, Baltimore, MD, United Statesb Schroeder Statistical Consulting, LLC, Ellicott City, MD, United Statesc Department of Pediatrics, Johns Hopkins School of Medicine, Baltimore, MD, United States

⁎ Corresponding author at: 6501 North Charles St., Balt938 4359; fax: +1 410 938 4364.

E-mail address: [email protected] (F. Dick

http://dx.doi.org/10.1016/j.schres.2014.04.0190920-9964/© 2014 Published by Elsevier B.V.

Please cite this article as: Dickerson, F., et alSchizophr. Res. (2014), http://dx.doi.org/10.

a b s t r a c t
a r t i c l e i n f o
Article history:
Received 9 February 2014Received in revised form 12 April 2014Accepted 17 April 2014Available online xxxx
Keywords:CognitiveSchizophreniaLongitudinal

Background: Cognitive deficits are a central feature of schizophrenia but it is not certain how cognitive function-ing changes over time. The purpose of this prospective longitudinal studywas to determine the temporal changeof cognitive functioning and the predictors of cognitive performance from among demographic, clinical, andbiological variables.Methods: Participants were individuals with schizophrenia or schizoaffective disorder whose cognitive function-ing was assessed at multiple time points with the Repeatable Battery for the Assessment of NeuropsychologicalStatus (RBANS). At the baseline visit participants had a blood sample drawn from which C-reactive protein,antibodies to Herpes Simplex Virus type 1, and selected genetic polymorphisms were measured. Repeatedmeasures linear regression was used to determine whether cognitive measures changed over time and which
variables predicted cognitive performance.Results: The sample consisted of 132 participants, mean age 43.7 years at baseline, who received a median of 3cognitive assessments over a period averaging 2.8 years. The RBANS Total score and Language index showedno statistically significant temporal change; performance on two indices, Immediate Memory and Attention,showed modest but statistically significant improvements (gains of 0.89 ± 0.33 and 0.76 ± 0.29 points peryear, respectively); Visuospatial/Constructional performance showed a modest but statistically significantdecline (of 0.80±0.25 points per year). Fewvariables predicted cognitive performance; however greater psychi-atric symptom severity was associated with worse cognitive performance for most cognitive measures.Conclusions: Cognitive functioning in middle-aged persons with schizophrenia showed an absence of decline formost measures and modest gains in some measures.
© 2014 Published by Elsevier B.V.

1. Introduction

Cognitive deficits are a central feature of schizophrenia and contributeto the social disability associatedwith thedisorder and to the highburdenof the disease (Green, 1996; Dickerson et al., 1999). As a group, personswith schizophrenia perform more poorly than age-matched controls onmeasures of verbal memory, executive functioning, attention, and pro-cessing speed (Heinrichs and Zakzanis, 1998). It is now established thatthese deficits are present at the time of illness onset and remain charac-teristic of schizophrenia over the illness course (Kurtz, 2005; Napalet al., 2012; Bora and Murray, 2013; Rodriguez-Sanchez et al., 2013).

imore, MD 21204. Tel.: +1 410

erson).

., A longitudinal study of cogn1016/j.schres.2014.04.019

It is less certain how cognitive functioning in this population chang-es over time. Some cross- sectional studies suggest that cognitive de-cline progresses over the illness course beyond that expected withnormal aging (Bilder et al., 1992; Sponheim et al., 2010). However,data from longitudinal studies suggest that cognitive function is stablyimpaired in individuals with schizophrenia who are assessed initiallynear the time of the first episode through the period of middle age(Nayak Savla et al., 2006; Zipursky et al., 2013) and that cognitive func-tioning may even improve (Szoke et al., 2008; Bonner-Jackson et al.,2010). On the other hand, studies in old age suggest accelerated cogni-tive deterioration, at least in a subgroup of patients with schizophrenia(Friedman et al., 1999; Harvey et al., 1999).

Results of longitudinal studies are not consistent about which vari-ables are associated with cognitive functioning in schizophrenia andwith cognitive change over time. In general, demographic variablessuch as age, gender and race have not been found to have a significanteffect in schizophrenia relative to controls (Heaton et al., 2001; Irani

itive functioning in schizophrenia: Clinical and biological predictors,

http://dx.doi.org/10.1016/j.schres.2014.04.019

mailto:[email protected]


http://www.sciencedirect.com/science/journal/09209964


2 F. Dickerson et al. / Schizophrenia Research xxx (2014) xxx–xxx

et al., 2011; Rajji et al., 2013). Within schizophrenia samples, moresevere psychiatric symptoms (Irani et al., 2011) and substance use(Heaton et al., 2001; Sponheim et al., 2010) at baseline have been asso-ciatedwithworse cognitive functioning as has the presence of psychosisat follow-up (Bonner-Jackson et al., 2010) but not all longitudinal stud-ies have included these variables. Longer duration of illness has beenassociated with more (Heaton et al., 2001; Sponheim et al., 2010), less(Irani et al., 2011), and nodifference (Klingberg et al., 2008) in cognitiveimpairment. Study findings are also not consistent about the effect ofbaseline cognitive level on subsequent cognitive performance (Heatonet al., 2001; Klingberg et al., 2008). Findings are also inconsistentabout the effect of medication, and the type of medication, on cognitivecourse (Klingberg et al., 2008; Szoke et al., 2008; Sponheim et al., 2010;Irani et al., 2011). Differences in patient samples, cognitive measures,and follow up intervals may account for these discrepancies amongstudies.

Some biological markers have been associated with cognitiveimpairment in cross-sectional studies of schizophrenia including bothenvironmental (Herpes Simplex Virus type 1 (HSV-1) seropositivity(Dickerson et al., 2003)), elevated C-reactive protein (CRP) (Dickersonet al., 2012) and genetic markers (COMT (Dickerson et al., 2007b))and LTA polymorphisms (Dickerson et al., 2007a). Similarly, co-morbid conditions such as diabetes (Dickinson et al., 2008) and ciga-rette smoking (Zhang et al., 2012; Morisano et al., 2013) have beenassociatedwith greater cognitive impairment in schizophrenia. Howev-er, previous longitudinal investigations have not studied the effect ofthese factors on the cognitive course.

The purpose of this prospective longitudinal study of non-geriatricadultswith schizophreniawas to determine the temporal change of cog-nitive functioning and the predictors of cognitive performance fromamong demographic, clinical, and biological variables. This study is thefirst to evaluate whether blood-based biological markers predict subse-quent cognitive decline in a sample of patients with schizophrenia.

2. Methods

Participants were individuals with a diagnosis of schizophrenia fromthe Stanley Research Program at Sheppard Pratt who were initially en-rolled in a study of the association between antibodies to infectiousagents and serious mental illness in the period between February 1,1999 and May 31, 2010. Participants met the following inclusioncriteria: 1) diagnosis of schizophrenia or schizoaffective disorder meet-ing criteria in the Diagnostic and Statistical Manual of Mental DisorderFourth Edition (DSM-IV, 1994) made by a board-certified psychiatristbased on the Structured Clinical Interview for Diagnosis for Axis I disor-ders (SCID) (First et al., 1996) and medical records; 2) age 18–65; 3)assessed at more than one time point through participation in theprogram's research studies. Exclusion criteria were: 1) history of intra-venous substance abuse; 2) history of mental retardation; 3) clinicallysignificant medical disorder that would affect cognitive performancesuch as epilepsy, HIV, history of encephalitis or head trauma, or anyother reported neurological disorder of the central nervous system; 4)primary diagnosis of substance abuse or dependence. Follow-up assess-ments were obtained during participants' enrollment in one or moreclinical trials performed by the same research group; in these trials par-ticipants in the current study were randomized to the add-on placebo,and not the add-on active, study medication. The clinical trials inwhich participants may have been enrolled were adjunctive trials ofomega-3 fatty acids (Fenton et al., 2001), azithromycin (Dickersonet al., 2009a), allopurinol (Dickerson et al., 2009b), or artemisinin(Dickerson et al., 2011).

All participants provided written informed consent for the initialstudy and the subsequent clinical trials in which they participated andall of the studies were approved by the Sheppard Pratt InstitutionalReview Board.

Please cite this article as: Dickerson, F., et al., A longitudinal study of cognSchizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.019

2.1. Measures

At time of the initial and the subsequent assessments, participantswere administered a brief cognitive test battery, the Repeatable Batteryfor the Assessment of Neuropsychological Status (RBANS) (Randolphet al., 1998). The RBANS was selected for the measurement of cognitivefunctioning because it is sensitive to the level and type of cognitiveimpairments in individuals with schizophrenia (Wilk et al., 2004).Test indices are Immediate Memory (comprised of List Learning andStory Memory subtests); Visuospatial/Constructional (comprised ofFigure Copy and Line Orientation subtests); Language (comprised ofPicture Naming and Fluency subtests); Attention (comprised of DigitSpan and Coding subtests); and Delayed Memory (comprised of ListRecall, Story Recall, Figure Recall, and List Recognition subtests). Eachindex score is expressed as an age-adjusted standard score with amean of approximately 100 and a standard deviation of approximately15 based on a normative study group of 540 healthy subjects, ranging inage from 20 to 89, matched to the U.S. Census on gender, ethnicity, andlevel of education. The index scores are combined to yield a Total RBANSscore which is a measure of overall cognitive functioning. Participantswere administered Form A of the RBANS at the initial assessment;on subsequent testing occasions Form A and Form B were [typically]counterbalanced.

At each occasion when the RBANS cognitive test was administrated,participants were interviewed and rated on the Positive and NegativeSyndrome Scale (PANSS); (Kay et al., 1987), a measure of psychiatricsymptom severity.

Each participant provided a blood sample at the time of the initialassessment from which C-Reactive Protein (CRP) levels and antibodiestoHerpes Simplex Virus type 1 (HSV-1)weremeasured as previously de-scribed (Dickerson et al., 2003). Blood sampleswere also analyzed for thelymphotoxin alpha (LTA), Cys13Arg polymorphism (rs2857713) and thevaline/methionine polymorphism of the catechol O-methyltransferasegene at amino acid 158 (COMT Val158Met polymorphism) as previouslydescribed (Dickerson et al., 2006). The CRP, HSV-1, LTA, and COMT vari-ableswere included in the current study because they have all been asso-ciated with cognitive functioning in schizophrenia in our prior studies.

At time of the initial assessment, participants were asked about de-mographic variables, their smoking status,whether they had a diagnosisof diabetes, and about their history of alcohol and drug abuse. Dataabout current medications were obtained from medical records.

2.2. Data analysis

Repeatedmeasures linear regression was used to determine whethercognitive measures changed over time and which variables predictedsubsequent cognitive performance. Variables included in the regressionmodels were the baseline variables of age (years), gender (M/F), race(Caucasian vs. other), yearsmaternal education, current cigarette smoker(Y/N), history of substance abuse (Y/N), age of illness onset (years), dura-tion of illness (years), and receipt of atypical anti-psychotic medicationsat baseline (Y/N). The following baseline biological variables were alsoincluded: CRP level, HSV-1 seropositivity status, LTA polymorphism(Arg/Arg, Arg/Cys, Cys/Cys), COMT polymorphism (Met/Met, Val/Met,Val/Val). In addition, the PANSS total symptom score at the time of eachcognitive assessment was included as a time-varying covariate.

Weighted regression was used to give more weight to observationsfurther away from the initial assessment: a weight of 0.1 was given tofollow-up intervals less than six months, a weight of 0.2 was given tofollow-up intervals from 6 months to two years, a weight of 0.3 wasgiven to follow-up intervals from 2 to 5 years, and a weight of 0.4 wasgiven to follow-up intervals greater than 5 years.

For each of the RBANS Total and Index scores, a weighted linear re-gression model was run that consisted of a single independent variable,duration of time elapsed since the first assessment (in years), in order todetermine which cognitive measures showed a statistically significant



3F. Dickerson et al. / Schizophrenia Research xxx (2014) xxx–xxx

change over time. In order to verify that observed changes in cognitivedomains are not artifacts of age-adjusted scoring, the same modelswere also run for the subtest scores on the RBANS since these scoresare not age-adjusted. The regression models for subtest scores wererun with and without age included as a covariate.

For the total and index measures that showed a statistically signifi-cant temporal change, an interaction model was run for each predictorvariable that consisted of that variable, plus time (years), plus the inter-action between the two. For measures that did not show a statisticallysignificant temporal change, amodelwas run for each predictor variablethat consisted of that variable as the only independent variable. Regres-sion results are reported as a regression coefficient plus or minus itsstandard error, plus a p-value indicating whether the coefficient is sig-nificantly different from zero. Analyses were done using R statisticalsoftware (version 2.13.2).

3. Results

3.1. The sample

The sample consisted of 132 participants of whom 61 (46.2%)received 2 assessments, 41 (31.1%) received 3 assessments, 19 (14.4%)received 4 assessments, 7 (5.3%) received 5 assessments, and 4 (3.0%)received 6 assessments over a period of 2.8 years on average (range19 days–10.2 years). Characteristics of the sample at time of the initialassessment are shown in Table 1.

At baseline, a total of 114 (86.4%) were receiving atypical antipsy-chotics including Clozapine (36, 27.3%), Olanzapine (35, 26.5%), Risper-idone (36, 27.3%), Quetiapine (15, 11.4%), Ziprasidone (6, 4.5%), andAripiprazole (4, 3.0%). A total of 65 (49.2%) were receiving anti-

Table 1Clinical characteristics of the sample at baseline (N = 132).

Mean ± s.d.or number (percent)

Demographic variablesAge at entry, years 43.7 ± 9.7White race 91 (68.9%)Female gender 52 (39.4%)Maternal education, years 12.3 ± 2.7

Clinical variablesAge at illness onset, years 21.0 ± 7.5Duration of illness, years 22.5 ± 10.0PANSS total symptom score 69.6 ± 11.7Receiving atypical antipsychotic 114 (86.4%)Current cigarette smoker 88 (66.7%)History of drug or alcohol abuse 71 (53.8%)Diabetes 27 (20.5%)

Biological variablesHigh sensitivity CRP level, mean 1.8 ± 1.7HSV-1 seropositivity 74 (56.5%)LTA polymorphisma

Arg/Arg 72 (58.1%)Arg/Cys 45 (36.3%)Cys/Cys 7 (5.6%)

COMT polymorphismb

Met/Met 26 (20.3%)Val/Met 60 (46.9%)Val/Val 42 (32.8%)

RBANS cognitive scoresTotal 67.9 ± 13.9Immediate Memory Index 68.4 ± 18.4Language Index 85.1 ± 13.8Attention Index 72.5 ± 15.8Delayed Memory Index 70.7 ± 19.0Visuospatial/Constructional Index 74.1 ± 17.9

a N = 124.b N = 128.


depressant medications, 44 (33.3%) anticholingeric medications, 14(10.6%) lithium, and 49 (37.1%) anticonvulsant mood stabilizers.

3.2. Temporal changes in cognitive variables

As shown in Table 2, theRBANSTotal score showedno statistically sig-nificant temporal change, but three out of five cognitive RBANS Indexscores showed modest but statistically significant changes over time.Two of these index scores increased: Immediate Memory (0.89 ± 0.33,p = 0.0075) and Attention (0.76 ± 0.29, p = 0.0091). Visuospatial/Constructional decreased modestly over time (−0.80 ± 0.25, p =0.0017). Delayed Memory showed a smaller increase that approachedstatistical significance (0.55 ± 0.30, p = 0.064). The Language indexshowed no statistically significant temporal change. Fig. 1 shows therelative magnitude of these temporal trends graphically. The averageincrease of 0.89 RBANS points per year for ImmediateMemory amountsto an overall increase of approximately nine points over a ten-yearperiod. The increase for Attention is slightly less: 0.76 points per yearon average or almost eight points in ten years. The decrease of 0.80points per year for Visuospatial/Constructional amounts to an eight-point decrease in ten years.

The RBANS subtests were also examined for temporal change, inorder to determine whether the temporal trends observed for theindex scores were artifacts of age-adjusted scoring. The results ofthe subtest analysis were largely consistent with those for theindex scores, suggesting that the observed changes noted in the pre-vious paragraph were not artifactual. List Learning, part of the Imme-diate Memory index, showed a statistically significant increase overtime, both with and without adjustment for age at baseline (0.40 ±0.12; 0.58 ± 0.13). Within the Attention index, which showed a signif-icant increase over time, the Coding subtest showed a significantincrease, but only with age-adjustment (0.48 ± 0.15). Within theVisuospatial/Constructional index which showed a significant decreaseover time, the Figure Copy subtest showed a significant decrease, bothwith and without age adjustment (−0.50 ± 0.07; −0.44 ± 0.07).The Language index score did not show a temporal change and neitherdid either of the component subtests. The Delayed Memory indexshowed a trend towards improvement over time; the List Recall andStory Recall subtests both showed a significant increase with adjust-ment for age (0.12 ± 0.05; 0.11 ± 0.05) but not without age adjust-ment. Also within the Delayed Memory index, the Figure Recalland Delayed Memory subtests both showed a significant decreaseover time without age adjustment (−0.24 ± 0.06; −0.20 ± 0,10) butwith age adjustment there was not a significant temporal change ineither of these measures.

3.3. Predictors of cognitive performance

A similar pattern of results was observed for the RBANS Total Scoreand the Language Index (Table 2). These measures showed no signifi-cant change over time, and both showed a positive association withwhite race and a negative associationwith PANSS total indicating higherpsychiatric symptom severity associated with lower cognitive perfor-mance. An additional predictor of RBANS Total score was maternaleducation.

Immediate Memory showed a statistically significant increase ofapproximately 0.89 points per year on average (Table 2). Higher scoreson the Immediate Memory Index were predicted by higher maternaleducation and lower PANSS total symptom score. In both of thesemodels, the effect of time was not statistically significant, and therewas no significant interaction with time. This means that the ability ofmaternal education and PANSS to predict ImmediateMemorywas inde-pendent of follow-up time.

The Attention index also showed a statistically significant in-crease over time (of approximately 0.76 points per year). Therewere significant interactions with time for three predictors: age at



Table 2Predictors of RBANS Total and Index scores from repeated measures regression models (N = 132).

Variable Total Immediate memory Delayed memory Language Attention Visuospat construc

Time – 0.89 ± 0.33 0.55 ± 0.30⁎ – 0.76 ± 0.29 −0.80 ± 0.25

DemographicsAge at entry – – – – a –

White race 6.50 ± 2.39 – – 4.84 ± 2.16 – 8.83 ± 2.90Female gender – – – – – –

Maternal educ 1.56 ± 0.41 2.17 ± 0.53 1.73 ± 0.60 – – 1.74 ± 0.50

Co-morbidityCig smoker – – – – – −7.91 ± 2.88Sub abuse – – – – – –

Diabetes – – – – – –

PsychiatricAge at onset – – – – – –

Duration – – – – a –

PANSS −0.15 ± 0.04 −0.24 ± 0.08 −0.23 ± 0.08 −0.13 ± 0.06 a –

Atypical meds – – – – – –

BiologicalCRP – – – – – –

HSV-1 – – a – – –

LTA – – – – – –

COMT – – a – – –

Numbers shown are parameter estimates ± standard errors for main effects only.Only statistically significant (p b 0.05) predictors of cognitive variables shown.⁎ p-Value for this term was p = 0.06.a Indicates variables with statistically significant interaction with time but no significant main effect.


evaluation (p = 0.0062), duration of illness (p = 0.0002), and PANSStotal symptom score (p = 0.018). To investigate the interactionsfurther, repeated measures regression models were run to test the ef-fect of time on Attention separately for high and low levels of durationof illness and baseline PANSS score, defined bymedian splits. For partic-ipants whose duration of illness was 23 years or less, therewas a highlysignificant increase in Attention score during the study period (1.51 ±0.33, p b 0.0001), but there was no significant temporal change inAttention for participants whose duration of illness was more than23 years (−0.52 ± 0.40, p = 0.20). However, there was no differencein Attention score change comparing participants who had low (≤69)

Fig. 1. RBANS Total and Index scores over the study period fromweighted regression analyses (The statistically significant slopes are solid lines; the one that is close to significant (p = 0.06)


vs. high (N69) baseline PANSS total symptom score, indicating no trueeffect modification.

The increase in the Delayed Memory index score of 0.55 points peryear came close to statistical significance (p = 0.06), so interactionmodels were run for this outcome. Two variables showed statisticallysignificant main effects and no interactions with time: maternal educa-tion and PANSS total symptom score. Two biologic variables showed in-teractions with time that were or approached statistical significance:HSV-1 (p = 0.058) and COMT polymorphism (Met/Met, Val/Met, orVal/Val; p = 0.02). To investigate these interactions further in theDelayed Memory domain, repeated measures regression models were

N= 132). RBANS= Repeatable Battery for the Assessment of Neuropsychological Status.is a dashed line; the ones that are not significant are dotted.



5F. Dickerson et al. / Schizophrenia Research xxx (2014) xxx–xxx

run to test the effect of time on Delayed Memory separately for eachlevel of these two variables. There was a statistically significant increasein DelayedMemory over time for participants whowere HSV-1 positive(1.09± 0.43, p= 0.012) but not for participants whowere HSV-1 neg-ative (−0.03 ± 0.40, p = 0.94). In spite of the statistically significantCOMT genotype by time interaction, there were no statistically signifi-cant time effects for any COMT subgroup (Met/Met, Val/Met, or Val/Val), indicating no true effect modification.

Visuospatial/Constructional was the only RBANS Index score toshow a statistically significant decrease. As was true for ImmediateMemory, interaction models showed statistically significant maineffects only, without any significant effects of time or interactionswith time. Both white race and higher maternal education predictedhigher Visuospatial/Constructional scores. Visuospatial/Constructionalwas also the only outcome to show a significant association witha co-morbid condition: being a cigarette smoker predicted lowerVisuospatial/Constructional scores. The ability of these three variablesto predict Visuospatial/Constructional score was independent of time.

4. Discussion

In this longitudinal study of outpatients with schizophrenia, wefound minimal changes in cognitive functioning over a period of up to10 years. These data indicate that for persons with schizophrenia whoare in middle age and have an illness duration averaging more than20 years, cognitive functioning is marked in general by stability andthe absence of deterioration. This finding is consistent with numerous,but not all, previous studies, and adds further weight to the conclusionthat, as a group, middle aged patients with schizophrenia do notevidence a decline in their cognitive functioning (Heaton et al., 2001;Nayak Savla et al., 2006; Bonner-Jackson et al., 2010; Zipursky et al.,2013).

While the total cognitive score did not show a significant changeover time in our cohort, we did find a statistically significant, albeitmodest, improvement in two of the domains, Immediate Memory andAttention, average gains of 0.89 and 0.76 points per year respectively.A previous meta-analysis of longitudinal studies of cognition in schizo-phrenia found improvement in verbal memory including in list learningand story memory which are similar to the subtests of RBANS Immedi-ate Memory (Szoke et al., 2008). Results from this same meta-analysisalso indicate improvement over time in digit span, one of the subtestsof RBANS Attention. As noted by previous investigators, we cannotrule out that these observed improvements are due to practice effects,rather than some remediation of cognitive deficits (Goldberg et al.,2010; Napal et al., 2012).

We found a statistically significant butmodest decline inVisuospatial/Constructional abilities, an average annual decrease of 0.80 points peryear, measured on the RBANS by Figure Copy and Judgment of LineOrientation subtests. It is of note that the Szoke et al. meta-analysisfound an absence of improvement over time on a task similar to theRBANS Figure Copy subtest. It is also of note that a longitudinal study ofpersons observed from pre-onset to post-onset schizophrenia found thelargest cognitive decline in visual spatial and motor performance(Meier et al., 2014); however, this study did not provide data aboutchanges taking place exclusively in the period after illness onset.

Because the index scores on the RBANS are age-adjusted, we lookedat change on the RBANS subtests, which are not adjusted for age. Theresults for the subtests were largely consistent with those of the indexscores, suggesting that the observed temporal changes in some of thecognitive domains were not artifacts of age-adjusted scoring.

Because there was little cognitive change over time for our cohort asa whole, there were few predictors of change. Herpes Simplex Virustype one (HSV-1) serostatuswas associatedwith change on theDelayedMemory score with individuals who were seropositive at baseline,showing an increase in performance. This finding is consistent withprevious studies in which we have shown an association between


seropositivity and cognitive performance (Dickerson et al., 2003,2012) and alsowhat is known about the pattern of latency and reactiva-tion of HSV-1 infection. It is possible that individuals who are seroposi-tivemay undergo alterations in their performance during the process ofviral reactivation and immune activation; however little is known aboutthe course of cognitive performance in individuals with HSV-1 so thisinterpretation is speculative. We did not find an overall main effect ofHSV-1 as we have in previous studies (Dickerson et al., 2003, 2012),which may be due to the relatively smaller sample size in this study.

We also found that a shorter duration of schizophrenia illness pre-dicted improvement on the Attention scale. This finding is consistentwith other studies, both longitudinal and cross sectional comparisons,which have found a shorter duration of illness to be associated withgreater cognitive improvement (Heaton et al., 2001; Sponheim et al.,2010). Our patients were ill on average more than 20 years, longerthan in these other studies suggesting that this association may alsobe found over an extended illness history.

The variables associated with cognitive performance as maineffects, not as predictors of change, are among those that would beexpected and are consistent with previous studies: higher maternaleducation, a proxy for higher premorbid socioeconomic status;white race; and lower PANSS score, indicating less severe psychiatricsymptoms (Bonner-Jackson et al., 2010; Irani et al., 2011).

It is also of note that we found adverse cognitive effects of ciga-rette smoking in that smokers performed worse on Visuospatial/Constructional. This finding replicates the results of Zhang et al.(2012) who also found that individuals with schizophrenia who werecigarette smokers performed worse than non-smokers on RBANSVisuospatial/Constructional; in the Zhang study, smokers also per-formed worse on RBANS Immediate Memory and on the Total score.Our results counter a prevailing assumption that cigarette smokinghas cognitive benefits for individuals with schizophrenia and serves toremediate deficits which are inherent in schizophrenia (Leonard et al.,2007). However, the relationship between cigarette smoking and cogni-tion in schizophrenia is complex and there may be confoundingvariables not accounted for in our analysis.

We also did not find a main effect for type of antipsychotic medica-tion, atypical vs. typical. Some previous longitudinal studies have alsonot found an effect of medication on cognitive course (Klingberg et al.,2008). However, it is of note that our data about patients' medicationswas limited to the time of the initial assessment; our analysis does notaccount for the complexity of medication regimens, or changes inthese regimens, over the time period of the study.

Limitations of our study include that the cognitive battery used wasrelatively brief; on the other hand, the battery iswell suited for repeatedassessments because of the alternate forms available and the ease ofadministration. In addition, we did not have a comparison group ofhealthy controls with which to compare cognitive stability with thatof the schizophrenia group. Thirdly, participants were, on average, mid-dle aged at baseline so our results do not inform the issue of whetherthere are changes in cognitive functioning closer to the time of illnessonset or in old age. Finally, we did not include all variables that maybe time-varying and affect cognitive status. As noted above, medicationtypes, doses, and combinations may impact cognition in ways that ourstudy was not able to detect.

Strengths of our study include that we evaluated a community-basedsample drawn from routine clinical settings. Importantly, our study in-cluded blood-based biological markers which, to our knowledge, is thefirst longitudinal study of cognition in schizophrenia to do so. Future stud-ies of the cognitive course in schizophrenia would benefit from a furtherintegration of data about genetic and environmental markers.

Role of the funding source

The funding source had no role in the collection of the data, theinterpretation of the data, or the preparation of the manuscript.




ContributorsAuthors Dickerson, Stallings, Origoni, and Yolken designed the study and wrote the

protocol. Authors Katsafanas, Schweinfurth, Savage, and Stallings contributed to subjectrecruitment and assessment. Author Origoni oversaw the dataset. Author Khushalanisupervised the diagnostic assessment of participants. Author Yolken performed the labo-ratory tests to generate the biological markers. Author Schroeder performed the statisticalanalysis. Author Dickerson wrote the first draft of the manuscript. All authors contributedto and have approved the final manuscript.

Conflict of interestDr. Yolken is a member of the Stanley Medical Research Institute Board of Directors

and Scientific Advisory Board. The terms of this arrangement are being managed by theJohns Hopkins University in accordance with its conflict of interest policies. The otherauthors declare that there are no conflicts of interest in relation to the subject of this study.

AcknowledgmentThis work was supported by grant 07R-1690 from the Stanley Medical Research

Institute.

References

Bilder, R.M., Lipschutz-Broch, L., Reiter, G., Geisler, S.H., Mayerhoff, D.I., Lieberman, J.A.,1992. Intellectual deficits in first-episode schizophrenia: evidence for progressivedeterioration. Schizophr. Bull. 18 (3), 437–448.

Bonner-Jackson, A., Grossman, L.S., Harrow, M., Rosen, C., 2010. Neurocognition in schizo-phrenia: a 20-year multi-follow-up of the course of processing speed and storedknowledge. Compr. Psychiatry 51 (5), 471–479.

Bora, E., Murray, R.M., 2013. Meta-analysis of cognitive deficits in ultra-high risk topsychosis and first-episode psychosis: do the cognitive deficits progress over, orafter, the onset of psychosis? Schizophr. Bull. (Jun 14, Epub ahead of print).

Dickerson, F., Boronow, J.J., Ringel, N., Parente, F., 1999. Social functioning andneurocognitive deficits in outpatients with schizophrenia: a 2-year follow-up.Schizophr. Res. 37 (1), 13–20.

Dickerson, F.B., Boronow, J.J., Stallings, C., Origoni, A.E., Ruslanova, I., Yolken, R.H., 2003.Association of serum antibodies to herpes simplex virus 1 with cognitive deficits inindividuals with schizophrenia. Arch. Gen. Psychiatry 60 (5), 466–472.

Dickerson, F.B., Boronow, J.J., Stallings, C., Origoni, A.E., Cole, S., Leister, F., Krivogorsky, B.,Yolken, R.H., 2006. The catechol O-methyltransferase Val158Met polymorphism andherpes simplex virus type 1 infection are risk factors for cognitive impairment inbipolar disorder: additive gene-environmental effects in a complex human psychiatricdisorder. Bipolar Disord. 8 (2), 124–132.

Dickerson, F., Boronow, J., Stallings, C., Origoni, A., Yolken, R., 2007a. The lymphotoxinCys13Arg polymorphism and cognitive functioning in individuals with schizophrenia.Schizophr. Res. 89 (1–3), 173–176.

Dickerson, F.B., Boronow, J.J., Stallings, C., Origoni, A.E., Sullens, A., Yolken, R.H., 2007b. Thecatechol O-methyltransferase Val158Met polymorphism is not associated withbroad-based cognitive functioning in schizophrenia. Schizophr. Res. 96 (1–3), 87–92.

Dickerson, F.B., Stallings, C.R., Boronow, J.J., Origoni, A.E., Yolken, R.H., 2009a. A double-blind trial of adjunctive azithromycin in individuals with schizophrenia who areseropositive for Toxoplasma gondii. Schizophr. Res. 112 (1–3), 198–199.

Dickerson, F.B., Stallings, C.R., Origoni, A.E., Sullens, A., Khushalani, S., Sandson, N., Yolken,R.H., 2009b. A double-blind trial of adjunctive allopurinol for schizophrenia.Schizophr. Res. 109 (1–3), 66–69.

Dickerson, F., Stallings, C., Vaughan, C., Origoni, A., Goga, J., Khushalani, S., Yolken, R., 2011.Artemisinin reduces the level of antibodies to gliadin in schizophrenia. Schizophr.Res. 129 (2–3), 196–200.

Dickerson, F., Stallings, C., Origoni, A., Vaughan, C., Khushalani, S., Yolken, R., 2012. Addi-tive effects of elevated C-reactive protein and exposure to Herpes Simplex Virus type1 on cognitive impairment in individuals with schizophrenia. Schizophr. Res. 134 (1),83–88.

Dickinson, D., Gold, J.M., Dickerson, F.B.,Medoff, D., Dixon, L.B., 2008. Evidence of exacerbatedcognitive deficits in schizophrenia patients with comorbid diabetes. Psychosomatics 49(2), 123–131.

Fenton, W.S., Dickerson, F., Boronow, J., Hibbeln, J.R., Knable, M., 2001. A placebo-controlled trial of omega-3 fatty acid (ethyl eicosapentaenoic acid) supplementationfor residual symptoms and cognitive impairment in schizophrenia. Am. J. Psychiatry158 (12), 2071–2074.


First, M., Gibbon,M., Spitzer, R.L., Williams, J.B.W., 1996. User's Guide for the SCID-I, Struc-tured Clinical Interview for DSM IV Axis I Disorders. Biometrics Research, New York,NY.

Friedman, J.I., Harvey, P.D., Kemether, E., Byne,W., Davis, K.L., 1999. Cognitive and functionalchanges with aging in schizophrenia. Biol. Psychiatry 46 (7), 921–928.

Goldberg, T.E., Keefe, R.S., Goldman, R.S., Robinson, D.G., Harvey, P.D., 2010. Circum-stances under which practice does not make perfect: a review of the practiceeffect literature in schizophrenia and its relevance to clinical treatment studies.Neuropsychopharmacology 35 (5), 1053–1062.

Green, M.F., 1996. What are the functional consequences of neurocognitive deficits inschizophrenia? Am. J. Psychiatry 153 (3), 321–330.

Harvey, P.D., Silverman, J.M., Mohs, R.C., Parrella, M., White, L., Powchik, P., Davidson, M.,Davis, K.L., 1999. Cognitive decline in late-life schizophrenia: a longitudinal study ofgeriatric chronically hospitalized patients. Biol. Psychiatry 45 (1), 32–40.

Heaton, R.K., Gladsjo, J.A., Palmer, B.W., Kuck, J., Marcotte, T.D., Jeste, D.V., 2001. Stabilityand course of neuropsychological deficits in schizophrenia. Arch. Gen. Psychiatry 58(1), 24–32.

Heinrichs, R.W., Zakzanis, K.K., 1998. Neurocognitive deficit in schizophrenia: a quantitativereview of the evidence. Neuropsychology 12 (3), 426–445.

Irani, F., Kalkstein, S., Moberg, E.A., Moberg, P.J., 2011. Neuropsychological performance inolder patients with schizophrenia: ameta-analysis of cross-sectional and longitudinalstudies. Schizophr. Bull. 37 (6), 1318–1326.

Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS)for schizophrenia. Schizophr. Bull. 13 (2), 261–276.

Klingberg, S., Wittorf, A., Sickinger, S., Buchkremer, G., Wiedemann, G., 2008. Course ofcognitive functioning during the stabilization phase of schizophrenia. J. Psychiatr.Res. 42 (4), 259–267.

Kurtz, M.M., 2005. Neurocognitive impairment across the lifespan in schizophrenia: anupdate. Schizophr. Res. 74 (1), 15–26.

Leonard, S., Mexal, S., Freedman, R., 2007. Smoking, genetics and schizophrenia: evidencefor self medication. J. Dual Diagn. 3 (3–4), 43–59.

Meier, M.H., Caspi, A., Reichenberg, A., Keefe, R.S., Fisher, H.L., Harrington, H., Houts, R.,Poulton, R., Moffitt, T.E., 2014. Neuropsychological decline in schizophrenia fromthe premorbid to the postonset period: evidence from a population-representativelongitudinal study. Am. J. Psychiatry 171 (1), 91–101.

Morisano, D., Wing, V.C., Sacco, K.A., Arenovich, T., George, T.P., 2013. Effects of tobaccosmoking on neuropsychological function in schizophrenia in comparison to otherpsychiatric disorders and non-psychiatric controls. Am. J. Addict. 22 (1), 46–53.

Napal, O., Ojeda, N., Sanchez, P., Elizagarate, E., Pena, J., Ezcurra, J., Gutierrez, M., 2012. Thecourse of the schizophrenia and its impact on cognition: a review of literature. ActasEsp. Psiquiatr. 40 (4), 198–220.

Nayak Savla, G., Moore, D.J., Roesch, S.C., Heaton, R.K., Jeste, D.V., Palmer, B.W., 2006. Anevaluation of longitudinal neurocognitive performance among middle-aged andolder schizophrenia patients: use of mixed-model analyses. Schizophr. Res. 83(2–3), 215–223.

Rajji, T.K., Voineskos, A.N., Butters, M.A., Miranda, D., Arenovich, T., Menon, M., Ismail, Z.,Kern, R.S., Mulsant, B.H., 2013. Cognitive performance of individuals with schizophreniaacross seven decades: a study using the MATRICS consensus cognitive battery. Am. J.Geriatr. Psychiatry 21 (2), 108–118.

Randolph, C., Tierney, M.C., Mohr, E., Chase, T.N., 1998. The Repeatable Battery for theAssessment of Neuropsychological Status (RBANS): preliminary clinical validity. J.Clin. Exp. Neuropsychol. 20 (3), 310–319.

Rodriguez-Sanchez, J.M., Ayesa-Arriola, R., Perez-Iglesias, R., Perianez, J.A.,Martinez-Garcia,O., Gomez-Ruiz, E., Tabares-Seisdedos, R., Crespo-Facorro, B., 2013. Course of cognitivedeficits infirst episode of non-affective psychosis: a 3-year follow-up study. Schizophr.Res. 150 (1), 121–128.

Sponheim, S.R., Jung, R.E., Seidman, L.J., Mesholam-Gately, R.I., Manoach, D.S., O'Leary, D.S.,Ho, B.C., Andreasen, N.C., Lauriello, J., Schulz, S.C., 2010. Cognitive deficits in recent-onset and chronic schizophrenia. J. Psychiatr. Res. 44 (7), 421–428.

Szoke, A., Trandafir, A., Dupont, M.E., Meary, A., Schurhoff, F., Leboyer, M., 2008. Longitudinalstudies of cognition in schizophrenia:meta-analysis. Br. J. Psychiatry J.Ment. Sci. 192 (4),248–257.

Wilk, C.M., Gold, J.M., Humber, K., Dickerson, F., Fenton, W.S., Buchanan, R.W., 2004.Brief cognitive assessment in schizophrenia: normative data for the RepeatableBattery for the Assessment of Neuropsychological Status. Schizophr. Res. 70(2–3), 175–186.

Zhang, X.Y., Chen da, C., Xiu,M.H., Haile, C.N., Sun, H., Lu, L., Kosten, T.A., Kosten, T.R., 2012.Cigarette smoking and cognitive function in Chinese male schizophrenia: a case-control study. PLoS One 7 (5), e36563.

Zipursky, R.B., Reilly, T.J., Murray, R.M., 2013. The myth of schizophrenia as a progressivebrain disease. Schizophr. Bull. 39 (6), 1363–1372.


http://refhub.elsevier.com/S0920-9964(14)00191-1/rf0005










































































































a longitudinal study of cognitive functioning in schizophrenia: clinical and biological predictors

Documents