Nordic Journal of Psychiatry

ISSN: 0803-9488 (Print) 1502-4725 (Online) Journal homepage: https://www.tandfonline.com/loi/ipsc20

Neurocognitive effects of atypical antipsychotics in

patients with first-episode schizophrenia

Yanyan Hou, Jiaheng Xie, Yanbo Yuan, Zhang Cheng, Xue Han, Lei Yang, Xin
Yu & Chuan Shi

To cite this article: Yanyan Hou, Jiaheng Xie, Yanbo Yuan, Zhang Cheng, Xue Han, Lei Yang,
Xin Yu & Chuan Shi (2020): Neurocognitive effects of atypical antipsychotics in patients with first-
episode schizophrenia, Nordic Journal of Psychiatry, DOI: 10.1080/08039488.2020.1771767

To link to this article: https://doi.org/10.1080/08039488.2020.1771767

Published online: 04 Jun 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ipsc20
NORDIC JOURNAL OF PSYCHIATRY
https://doi.org/10.1080/08039488.2020.1771767

ORIGINAL ARTICLE

Neurocognitive effects of atypical antipsychotics in patients with first-episode

schizophrenia
Yanyan Houa,b,c,d,e, Jiaheng Xief, Yanbo Yuana,b,c,d, Zhang Chenga,b,c,d, Xue Hana,b,c,d, Lei Yanga,b,c,d,
Xin Yua,b,c,d and Chuan Shia,b,c,d
a
Peking University Sixth Hospital, Beijing, 100191, China; bPeking University Institute of Mental Health, Beijing, 100191, China; cNHC Key
Laboratory of Mental Health, Ministry of Health (Peking University), Beijing, 100191, China; dThe National Clinical Research Center for
Mental Health Disorders (Peking University Sixth Hospital), Beijing, 100191, China; eQingdao Mental Health Center, Qingdao University,
Shandong, 266034, China; fPeking University, Beijing, 100191, China

ABSTRACT ARTICLE HISTORY

Introduction: Cognitive impairment is a core feature of schizophrenia. The effects of atypical antipsy- Received 2 November 2019
chotics on the cognitive functions of patients with first-episode schizophrenia have not been compre- Revised 13 March 2020
hensively investigated so far. This study aims to compare neurocognitive effects of risperidone, Accepted 17 May 2020
olanzapine, and aripiprazole for first-episode schizophrenia.
KEYWORDS
Methods: The study was a multicenter, randomized, open-label clinical trial. 546 patients were ran- Neurocognitive function;
domly divided into three medication groups, and followed up for 1 year. Cognitive performance was atypical antipsychotics; first-
evaluated with a neuropsychological test battery. The Clinical trials.gov ID of the study is episode schizophrenia
NCT01057849.
Results: At 6 months, treatment resulted in significant improvements in all three groups in most cogni-
tive domains except verbal learning and memory. At 12 months, three treatment groups had further
improvements in three cognitive domains, but visual learning and memory performance dropped
back to baseline.
Conclusion: All three atypical antipsychotics tested in the study can potentially improve cognitive per-
formance in first-episode schizophrenia, but no significant difference in the degree of improvement
was found between drugs.

Introduction have a wider range of effects on neurocognitive function and

more benefits for improving cognition [8,9]. A meta-analysis
Cognitive impairment is one of the central clinical features of
showed that 19 out of 23 studies demonstrated atypical anti-
schizophrenia, which mainly involves seven domains: proc-
psychotics are significantly more effective in maintaining or
essing speed, attention/vigilance, working memory, vocabu-
improving neurocognitive function in patients with schizo-
lary learning, visual learning, reasoning, and problem-solving
phrenia [10]. However, previous studies have some limitations.
skills [1]. These impairments may appear prior to the onset
Most of them used typical antipsychotics as controls; the sub-
of obvious psychotic symptoms [2], and neurocognitive func-
jects, first-episode schizophrenia or chronic schizophrenia,
tion might has a different trajectory in the disease course as
were inconsistent. In addition, the sample sizes, various dur-
it has no strong correlation with the severity of psychotic
ation of drug treatment monitoring, application of adjuvant
symptoms [3,4]. However, psychotic episodes may further
drugs (e.g. anticholinergics and benzodiazepines), and hetero-
aggravate decline in cognitive performance. Cognitive defi-
geneity of cognitive tests may affect the observed results.
cits in schizophrenia may be mainly related to neurodevelop-
This study aims to compare cognitive effects of three
ment abnormalities. However, previous research has shown
atypical antipsychotics, i.e. risperidone, olanzapine, and aripi-
that the combination of metabolic syndromes and their
prazole in first-episode schizophrenia after 6 months and
related components (e.g. obesity and diabetes) may further
12 months of treatment.
contribute to cognitive impairment in schizophrenia [5].
Cognitive impairment is strongly associated with functional
outcomes [6,7], therefore, it is considered a potential direc- Methods
tion for psychopharmacological treatment.
Settings and design
Currently, the cognitive effects of atypical antipsychotic
drugs for patients with schizophrenia are relatively consistent: This study aims to compare the changes in cognitive per-
compared with typical antipsychotics, atypical antipsychotics formance before and after the three atypical antipsychotic

CONTACT Chuan Shi shichuan@bjmu.edu.cn; Xin Yu yuxin@bjmu.edu.cn Peking University Sixth Hospital, Peking University Institute of Mental
Health, Peking University, No. 51 Huayuanbei Road, Haidian District, Beijing 100191, China
ß 2020 The Nordic Psychiatric Association
2 Y. HOU ET AL.

treatments, and whether different medication provides differ- et al. [14]. A Chinese version of this assessment with high
ential effect. Background and study design were previously reliability and validity was created and further revised by Shi
described in detail [11]. This multicentre, randomized, open- et al. [15]. The 10 tests used in the study contained 16 varia-
label clinical trial was conducted from January 2008 to bles. These tests comprise six cognitive domains: speed of
December 2010 at six psychiatric centres in China. Patients processing (Trail Making Test, Part A, action naming, animal
were randomly divided into three groups treated with risper- naming, Color Trails Test 1, Stroop Color Test, and Stroop
idone, olanzapine, and aripiprazole, and then followed up for Word Test), vocabulary learning and memory (Hopkins Verbal
1 year. Learning Test-Revised), visual learning and memory (Brief
Visuospatial Memory Test-Revised), working memory and
attention (Wechsler Memory Scale-spatial span subtest and
Participants
paced auditory serial addition test), executive function (Color
Inclusion criteria were as follows: both in- and outpatients; Trails Test 2 and Stroop Color–Word Test), and fine motor
both genders; diagnosis of schizophrenia established with function (Grooved Pegboard Test).
Structured Clinical Interview for the DSM-IV Axis I Disorder, The raw scores for the 10 tests were converted into scaled
patient edition (SCID-I/P); age between 16 and 45; psychosis scores based on a mean of 10 and a SD of 3. Scaled scores
onset after age of 15; no previous systematic psychiatric were adjusted further by age, gender, education, and city of
treatment (continuous treatment for more than 1 month or birth and city of growth, and converted into the standardized
intermittent treatment for more than 3 months in total); t score with a mean of 50 and a SD of 10. Three domains
school education of more than 3 years. Exclusion criteria: his- (speed of processing, working memory and attention, execu-
tory or diagnosis of neurological disorders or other major tive function) were examined using more than one test or vari-
medical conditions, prolonged loss of consciousness (> 1 h), able, we used their summary scores, which were calculated by
intellectual disability; having participated in another clinical averaging the t scores for the contributing variables. The cog-
pharmacology study within one month before screening; nitive composite score was computed by averaging t scores of
alcohol and/or drug abuse or dependence within 3 months; six cognitive domains for each treatment group.
women during pregnancy or lactation; contraindication to
olanzapine, aripiprazole or risperidone; colour blindness;
hearing disability. Statistical analyses
This study was approved by the Ethics Committee of the SPSS 22.0 (SPSS IBM Inc., Chicago, IL) software was used for
Sixth Hospital of Peking University (Beijing, China). All statistical analysis. In the descriptive analysis, categorical vari-
patients signed written informed consent forms before ables were described as percentage and continuous variables
enrollment, and have been registered as legislation requires, as mean and standard deviation. In hypothesis testing, chi-
the clinical trials.gov ID of the study is NCT01057849. squared test was used for categorical variables; for continu-
ous variables that follow the normal distribution, one-way
Intervention ANOVA and repeated measures ANOVA were applied; for
continuous variables that do not follow the normal distribu-
All patients started antipsychotic monotherapy; antipsychotics tion, Kruskal-Wallis H test and Spearman correlation test
doses started off at 3-mg risperidone, 15–30 mg aripiprazole or were applied; the addition of anticholinergic treatment dur-
10–20 mg olanzapine per day and increased to a standard ing the first 6 months was examined using ANCOVA and
therapeutic range within two weeks. During the study, no Spearman correlation, a ¼ 0.05.
other antipsychotics were used; if necessary, small doses of
benzodiazepines (0.5–1.5 mg oral lorazepam or 2–4 mg oral or
im clonazepam per day) could be prescribed for agitation and Results
oral benzhexol (2–6 mg per day) or promethazine (25–75 mg
Demographic and clinical characteristics
per day) for extrapyramidal side effects (EPS).
546 patients were randomly divided into the risperidone
(n ¼ 189), olanzapine (n ¼ 178), and aripiprazole (n ¼ 179)
Instruments
groups (Fig. 1). The number of patients who completed base-
Mental state was assessed by trained psychiatrists at baseline line cognitive and clinical assessments were 498, 209 of 498
and 4-week, 8-week, and 13-week and then in 3-month inter- patients completed the 6 months cognitive tests and did not
vals until the end of 1-year follow-up. Symptoms were change the antipsychotics; finally, 125 patients completed all
assessed by the Positive and Negative Syndrome Scale assessments including the first 6 month and 12 month. Those
(PANSS) [12]. Adverse events were assessed using the Udvalg 125 patients were eventually included in the analysis of
for kliniske undersogelser (UKU) [13] side effect scale. this study.
Cognitive performance was evaluated at baseline, 6- Demographics and medication use about the 125 patients
month and endpoint with a neuropsychological test battery are presented in Table 1. No significant difference in age, sex,
adapted from MATRICS Consensus Cognitive Battery (MCCB), education level, duration, body mass index (BMI), PANSS score,
which consists of 10 tests selected from more than 90 nomi- or cognitive composite score was found among treatment
nated tests. The US norm of MCCB was developed by Kern groups at baseline; the use of benzodiazepines or
 NORDIC JOURNAL OF PSYCHIATRY 3

583 paents assessed for eligibility

37 ineligible

546 paents randomized

189 assigned to risperidone 178 assigned to olanzapine 179 assigned to aripiprazole

172 completed baseline assessment 157 completed baseline assessment 169 completed baseline assessment

93 dropped out 85 dropped out 105dropped out

43 withdrawn 37 withdrawn 35 withdrawn
18 insufficient efficacy 6 insufficient efficacy 45 insufficient efficacy
13 side effects 15 side effects 0 side effects
19 other reasons 27 other reasons 25 other reasons

79 completed the month 6 follow-up 72 completed the month 6 follow-up 64 completed the month 6 follow-up

32 dropped out 37 dropped out 21 dropped out

10 withdrawn 16 withdrawn 10 withdrawn

4 insufficient efficacy 0 insufficient efficacy 2 insufficient efficacy

2 side effects 7 side effects 1 side effects

16 other reasons 14 other reasons 8 other reasons

47 completed the month 12 follow-up 35 completed the month 12 follow-up 43 completed the month 12 follow-up

Figure 1. Trial profile.

promethazine during the 1-year follow-up period is also no 12 months are presented in Table 2. Repeated measures
difference between treatment groups. There was a significant ANOVA was used to assess the effect of medication group
difference in anticholinergic use (v2 ¼ 14.411, p ¼ 0.001): after on patients’ test scores. As anticholinergic use was different,
Bonferroni correction for multiple comparisons, patients from the combination of anticholinergic as a covariate was
olanzapine group (20.0%) used less anticholinergic than aripi- included in the analysis. Different medication groups had no
prazole group (62.8%), but there is no significant difference significant effect on patients’ overall cognitive performance
between risperidone and aripiprazole or olanzapine group. or any specific domain (p > 0.05). Interactions between time
Baseline characteristics for the 125 patients in final sample points and medication groups had no significant effect on
were compared with the 373 patients who had baseline cog- cognitive test performance (p > 0.05). At different time points
nitive tests and did not complete month 6 and month 12 after treatment started, various cognitive performances
tests using chi-square tests or t tests. The results show that emerged in the domains of speed of processing, visual learn-
patients in final sample had higher education（ t ¼ 2.55, ing and memory, working memory and attention, executive
p ¼ 0.011） and lower BMI(t ¼ 2.10, p ¼ 0.036). The other function, and fine motor function (p < 0.05).
measures, including PNASS scores and cognitive composite To further examine treatment time’s differential effect on
score at baseline were not different between groups. cognitive function, multiple comparisons were performed
using the Bonferroni test with time points as main effect
(Table 3), and the significance level was adjusted to 0.017 (¼
Changes in cognitive performance after treatment with
0.05/3). Results show, overall cognitive performance signifi-
atypical antipsychotics
cantly improved at 6 and 12 months compared to baseline,
Cognitive performance scores from each group after treat- but there is no difference between 12 and 6 months. For the
ment with atypical antipsychotics for 6 months and five domains that showed improvements in the previous
4 Y. HOU ET AL.

Table 1. Demographics, clinical characteristics, and medication at baseline.

Risperidone Olanzapine Aripiprazole
(n ¼ 47) Mean ± SD (n ¼ 35) Mean ± SD (n ¼ 43) Mean ± SD Total (n ¼ 125) Mean ± SD p-Value
Age (years) 24.60 ± 6.60 23.51 ± 5.66 25.37 ± 6.48 24.56 ± 6.30 0.456a
Gender, n (%) 45 (100) 34 (100) 38 (100) 117 (100) 0.975b
Male 20 (44.4) 15 (44.1) 16 (42.1) 51 (43.6)
Female 25 (55.6) 19 (55.9) 22 (57.9) 66 (56.4)
Education (years) 12.64 ± 2.57 12.83 ± 2.48 13.23 ± 2.51 12.90 ± 2.52 0.510a
Duration (months) 10.71 ± 10.74 12.00 ± 10.58 13.00 ± 13.74 11.46 ± 11.47 0.827a
BMI (kg/m2) 19.95 ± 5.14 20.16 ± 3.12 20.74 ± 3.15 20.57 ± 3.04 0.643a
Baseline PANSS
Positive symptoms 23.28 ± 4.84 23.49 ± 4.59 22.02 ± 5.06 22.90 ± 4.85 0.267a
Negative symptoms 22.00 ± 8.02 20.34 ± 7.08 22.26 ± 6.37 21.62 ± 7.21 0.461c
General 41.96 ± 6.91 40.66 ± 6.73 42.00 ± 8.01 41.60 ± 7.23 0.661c
psychopathology
Total 87.22 ± 13.06 84.49 ± 12.95 86.19 ± 14.13 86.14 ± 13.25 0.661c
Baseline cognitive 34.17 ± 8.84 35.19 ± 7.56 33.32 ± 8.26 34.18 ± 8.27 0.640c
composite score
Anticholinergic used, 47 (100) 35 (100) 43 (100) 125 (100) 0.001b
n (%)
Yes 20 (42.6) 7 (20.0) 27 (62.8) 54 (43.2)
No 27 (57.4) 28 (80.0) 16 (37.2) 71 (56.8)
Benzodiazepine used, 47 (100) 35 (100) 43 (100) 125 (100) 0.403b
n (%)
Yes 8 (17.0) 5 (14.3) 11 (25.6) 24 (19.2)
No 39 (83.0) 30 (85.7) 32 (74.4) 101 (80.8)
Promethazine used, 47 (100) 35 (100) 43 (100) 125 (100) 0.638b
n (%)
Yes 3 (6.4) 2 (5.7) 1 (2.3) 6 (4.8)
No 44 (93.6) 33 (94.3) 42 (97.7) 119 (95.2)
a
Kruskal-Wallis H test.
b
Chi-squared test.
c
One-way ANOVA.
d
Refers to use during the 1-year follow-up period.

Table 2. Cognitive scores at three time points by group.

Risperidone Olanzapine Aripiprazole
Cognitive function Time point Mean ± SD Mean ± SD Mean ± SD p-Valuea p-Valueb,c
Composite score Baseline 33.87 ± 8.70 35.34 ± 7.48 32.57 ± 8.47 0.698 <0.001
6 months 36.90 ± 6.74 37.44 ± 7.78 37.23 ± 5.49
12 months 37.45 ± 6.02 38.10 ± 7.59 36.85 ± 6.55
Speed of processing Baseline 35.81 ± 10.32 36.51 ± 9.65 34.81 ± 9.04 0.505 <0.001
6 months 39.26 ± 8.13 39.48 ± 7.89 37.70 ± 7.04
12 months 40.06 ± 7.83 42.10 ± 8.40 39.60 ± 7.47
Vocabulary learning and memory Baseline 36.55 ± 12.34 38.52 ± 10.79 33.59 ± 14.01 0.464 0.725
6 months 36.41 ± 10.59 36.12 ± 9.78 37.90 ± 11.21
12 months 35.20 ± 10.50 35.36 ± 12.37 36.03 ± 14.38
Visual learning and memory Baseline 37.54 ± 12.18 37.33 ± 12.86 37.14 ± 11.84 0.790 0.002
6 months 41.17 ± 11.11 39.76 ± 10.44 41.85 ± 8.85
12 months 36.71 ± 8.48 37.01 ± 9.81 38.72 ± 8.75
Working memory and attention Baseline 35.18 ± 10.00 35.32 ± 10.95 35.61 ± 8.73 0.716 <0.001
6 months 40.24 ± 8.37 40.94 ± 11.82 41.61 ± 9.31
12 months 41.98 ± 8.67 42.03 ± 10.41 42.53 ± 8.35
Executive function Baseline 30.62 ± 12.52 32.32 ± 10.18 31.52 ± 10.54 0.909 <0.001
6 months 35.81 ± 10.33 36.85 ± 9.42 36.29 ± 7.91
12 months 38.04 ± 8.90 37.93 ± 11.46 38.73 ± 9.16
Fine motor function Baseline 27.60 ± 8.71 26.79 ± 11.88 25.23 ± 10.82 0.862 <0.001
6 months 29.11 ± 9.06 30.63 ± 11.26 28.90 ± 9.18
12 months 33.28 ± 9.18 32.96 ± 10.79 31.07 ± 11.05
a
p-Value of medication group’s main effect.
b
p-Value of time point’s main effect.
c
All cognitive domain scores failed to pass Mauchly’s sphericity test (p < 0.05), p-values presented in this column have been Huynh-Feldt corrected.

analysis, speed of processing, executive function, and fine noteworthy, which means the treatment effect is likely not
motor function all had significant improvements between sustainable beyond 6 months.
6 months and baseline and between 12 and 6 months.
Compared to baseline scores, working memory and attention
Correlation between changes in clinical symptoms and
improved at 6 and 12 months, but there was no significant
cognitive performance
change between 12 and 6 months. For visual learning and
memory, improvement was significant at 6 months, but per- Table 4 shows results of the relationship between changes in
formance dropping back to baseline levels at 12 months is scores of clinical symptom and cognitive function from
 NORDIC JOURNAL OF PSYCHIATRY 5

Table 3. Medication effect on cognitive functions over time.

Time’s main effect on cognitive function
6 months – Baseline 12 months – Baseline 12 months – 6 months
Baseline 6 months 12 months (95% CI) (95%CI) (95%CI)
Cognitive function Mean ± SD Mean ± SD Mean ± SD 2.97 (1.35, 4.59) 3.47 (1.81, 5.12) 0.50 (0.56, 1.56)
Speed of processing 35.68 ± 9.66 38.80 ± 7.69 40.49 ± 7.88 2.67 (1.09, 4.24) 4.79 (3.19, 6.39) 2.12 (1.13, 3.12)
Vocabulary learning and memory 36.12 ± 12.57 36.82 ± 10.52 35.52 ± 12.31 0.18 (2.71, 3.08) 0.68 (3.95, 2.59) 0.86 (3.20, 1.47)
Visual learning and memory 37.34 ± 12.16 41.01 ± 10.14 37.49 ± 8.93 3.47 (0.54, 6.39) 0.035 (2.62, 2.69) 3.43 (5.80, 1.06)
Working memory and attention 35.36 ± 9.82 40.88 ± 9.69 42.47 ± 9.03 5.47 (3.34, 7.59) 6.86 (4.51, 9.20) 1.39 (0.62, 3.40)
Executive function 31.40 ± 11.18 36.26 ± 9.25 38.24 ± 9.68 4.34 (2.27, 6.41) 6.38 (4.05, 8.70) 2.03 (0.54, 3.53)
Fine motor function 26.56 ± 10.38 29.47 ± 9.72 32.44 ± 10.27 2.92 (0.93, 4.91) 5.94 (3.84, 8.05) 3.02 (1.20, 4.84)
p < 0.017, p < 0.001.

Table 4. Correlation between clinical symptoms and change in cognition at 6 months (Spearman’s rank-order correlation, n ¼ 125).
Vocabulary
Cognitive Speed of learning Visual learning Working memory Executive Fine motor
PANSS composite score processing and memory and memory and attention function function
Positive symptoms 0.089 0.150 0.147 0.036 0.074 0.008 0.088
Negative symptoms 0.228 0.180 0.166 0.172 0.189 0.096 0.210
General psychopathology 0.189 0.107 0.079 0.101 0.306 0.190 0.081
Total score 0.202 0.117 0.112 0.162 0.278 0.161 0.178
p < 0.05, p < 0.01.

baseline to 6 month: decreases in PANSS total score and score of vocabulary learning and memory domain improved
negative symptoms score both correlate with the increase in (mean [SE] change, 1.550 [1.798]), but the difference was not
cognitive composite score (r ¼ 0.202, p ¼ 0.036; r ¼ 0.228, significant (p ¼ 0.199).
p ¼ 0.017), which suggests cognitive functions improve as
symptoms alleviate. Additionally, PANSS scores correlate with
Discussions
a number of cognitive domains. Decrease in PANSS total
score correlates with the increase in working memory and This study compares cognitive effects of three atypical anti-
attention (r ¼ 0.278, p ¼ 0.002); Decrease in negative symp- psychotics, i.e. risperidone, olanzapine, and aripiprazole in
toms score correlates with the increase in speed of process- first-episode schizophrenia. Overall cognitive performance
ing, working memory and attention, and fine motor function improved after atypical antipsychotics treatment. Aside from
(r ¼ 180, p ¼ 0.045; r ¼ 189, p ¼ 0.039; r ¼ 210, p ¼ 0.019); vocabulary learning and memory, scores in all domains went
decrease in general psychopathology score correlates with up at 6 months, and improvement was sustained in most
the increase in working memory and attention, and execu- domains later in the study. Working memory and attention
tive function (r ¼ 0.306, p ¼ 0.001; r ¼ 0.190, p ¼ 0.036). had no further improvement and visual learning and mem-
ory performance dropped to baseline levels at 12 months.
Previous studies have shown that atypical antipsychotics
Association between the dose of anticholinergic
can compensate for cognitive impairment in schizophrenia,
medications and cognitive performance
and different types of atypical antipsychotics may have dif-
The addition of anticholinergic treatment during the first ferential effects on certain cognitive domains. Consensus
6 months was examined using ANCOVA and Spearman cor- findings on risperidone include the improvement of multiple
relation. A trend for significant negative association between cognitive domains, i.e. speed of processing, working memory,
the dose of anticholinergic medications and T score of executive function, attention, vocabulary learning and mem-
Hopkins Verbal Learning—total score (r ¼ 0.255, p ¼ 0.087), ory, and visuospatial memory [10,16]. Olanzapine can
Hopkins Verbal Learning Test— delayed recall（ r ¼ 0.289, improve speed of processing, vocabulary learning and mem-
p ¼ 0.060） , and Stroop Color/Word test (r ¼ 0.265, ory, attention/vigilance, and executive function [16,17].
p ¼ 0.075) was observed. In addition, a significant negative Relatively few studies have reported the effects of aripipra-
association was observed between the dose of anticholiner- zole, among those that have; some reported improvement in
gic medications and T score of vocabulary learning and reaction time and accuracy of working memory [18], motor
memory domain (r ¼ 0.305, p ¼ 0.047). There was no signifi- speed [19], and visual learning and memory [16]. In this
cant association between the dose of anticholinergic medica- study, all three medications significantly improved some cog-
tions and cognitive composite score. For the 43 patients who nitive domains, namely, speed of processing, working mem-
had anticholinergic medications added during the first ory and attention, executive function, and fine motor
6 months of treatment, T score of vocabulary learning and function, and also captured the short-term improvement on
memory domain worsened (mean [SE] change, 2.284 visual learning and memory that had been reversed in fur-
[2.359]) at the 6-month assessment; other patients whose T ther follow-up. These results were consistent with previous
6 Y. HOU ET AL.

findings, suggesting that atypical antipsychotics may improve Atypical antipsychotic drugs have been reported to
certain cognitive domains in early schizophrenia. However, improve cognitive function in schizophrenia, but the mech-
our findings differed in that these three atypical antipsy- anism of effect is still unclear. Cognitive symptoms are cur-
chotics did not improve the cognitive domain of vocabulary rently considered linked to abnormal information processing
learning and memory. The difference might be related to in the dorsolateral prefrontal cortex, while the prefrontal cor-
anticholinergic use—the high serum anticholinergic activities tex in schizophrenia is regulated by neurotransmitters dopa-
in the schizophrenic patients in the previous studies resulted mine, serotonin, acetylcholine, norepinephrine, glutamate,
in poor word recall [20,21], vocabulary learning and memory and histamine. Previous studies have shown that blockage of
[22,23], and overall cognitive performance [24]. In this study, the D1 receptor in the dorsolateral prefrontal cortex can dis-
there was a significant negative association between the rupt working memory [33], some D2 and D4 receptor ago-
dose of anticholinergic medications and T score of vocabu- nists can improve cognition by promoting cortical dopamine
lary learning and memory domain among these patients release [34], and selective D3 receptor antagonists help to
who had anticholinergic medication added during the first improve cognitive symptoms by increasing the concentration
6 months, which might explain the lack of improvement in of prefrontal dopamine and acetylcholine [35]. Partial 5-HT1A
this domain. This result does not suggest that anticholinergic receptor activation can alleviate cognitive symptoms; while
treatments should be withheld to aid cognition, however, 5-HT2A receptor antagonism increases the dopamine release
since the extrapyramidal symptoms that are the targets of through disinhibition also achieve similar effects. 5-HT2C
anticholinergic treatment are also associated with worse cog- antagonists reverse the inhibition of norepinephrine and the
nitive performance [25]. We should strengthen the manage- dopamine release, thereby improving cognition through
ment of the use of anticholinergic medications: assessing higher level of norepinephrine and dopamine function in the
extrapyramidal symptoms regularly; reducing the unneces- prefrontal cortex [36]. The central cholinergic system is
sary use of anticholinergic medications and shortening highly involved in learning and memory. Specifically, the
the duration. muscarinic M1 receptor is considered as most closely related
Dividing the 12 months study into two 6-month periods, to cognition. Pharmacological stimulation of muscarinic M1
two domains, visual learning and memory, and working receptors can enhance learning and memory. In addition,
memory and attention showed different treatment effects selective agonists of alpha7 nicotinic acetylcholine receptors
over the two periods. The significant improvement in visual increase the dopamine release in the cerebral cortex and
learning and memory during the first half of the study was improve learning and attention. The ionic N-methyl-D-aspar-
not sustained during the second half. Patients’ performance tate receptors (NMDARs) are closely related to synaptic plas-
went back to baseline levels. For working memory and atten- ticity and learning and attention [37]. NMDAR hypofunction
tion, treatment effect was sustained during second half but in schizophrenia causes the loss of complementary stimula-
did not get significantly larger. The lack of longer-term tory effects in the midbrain-cortical pathway, leading to low
improvement in these two domains might be in line with cortical and midbrain dopaminergic pathway activity, which
the overall unresponsiveness in vocabulary learning and may explain the cognitive impairment of schizophrenia [36].
memory, as these memory-related functions might be par- In addition, norepinephrine a2 receptor agonists can improve
ticularly difficult to treat. Structural and functional deficits in working memory, and histamine H3 receptor antagonists can
hippocampus had been consistently observed in patients improve cognition by increasing inter-synaptic hyaluronic
with schizophrenia and the changes started during first epi- acid (HA) concentration [33].
sode [26,27]. Many studies have shown that hippocampal Cognitive impairment in schizophrenia coexists with both
volume correlates with memory-related functions in psych- negative and disintegrative symptoms. Many studies have
otic patients [28,29]. If the medications can only lead to func- shown that reduction of negative symptoms is associated
tional but not structural improvements, the limitation of with improvement of cognitive function, and the severity of
therapeutic effects may be the result of ceiling effect from positive symptoms is unrelated to this effect [38–40]. The
hippocampal deficits. results of this study are consistent with previous findings in
No significant difference in cognitive improvements was showing that reduction of negative symptoms significantly
found among the three antipsychotics, which confirms previ- relates to enhanced overall cognitive function and a few
ous findings that suggest similar effects of atypical antipsy- domains. In addition, reduction of general psychopatho-
chotics on cognition [30,31]. logical symptoms is associated with improvement of working
This study did not include a negative control group. memory and attention and executive function, with a par-
Therefore, the observed improvement might be in question ticularly significant relationship between the remission of
for potential placebo and practice effect. A study conducted attention disorder and the improvement of working memory
by Keefe and colleagues [32]specifically tested placebo and attention. Moreover, reduction of overall symptoms
response and practice effects of MCCB in clinical trials of score indicated that the patients were less affected by symp-
schizophrenia: the treatment effect in the placebo group was toms of schizophrenia and could better focus on cognitive
approximately 2 t-score points in composite scores, and the tasks, which may be the cause of the improvement in atten-
effect size was very small. In comparison, it is reasonable to tion and working memory.
assume the cognitive improvement observed in this study is Limitations of this study included (1) the absence of drug-
by large attributable to medications. free control group, which means cognition improvement
 NORDIC JOURNAL OF PSYCHIATRY 7

could be a mere practice effect. (2) the relatively small sam- [8] Bilder RM, Goldman RS, Volavka J, et al. Neurocognitive effects of
ple size, which did not allow us to conduct assessments at clozapine, olanzapine, risperidone, and haloperidol in patients
with chronic schizophrenia or schizoaffective disorder. Am J
multiple follow-up time points and to understand the trajec-
Psychiatry. 2002;159(6):1018–1028.
tory of cognitive improvement; (3) that lack of neuroimaging [9] Purdon SE, Jones BD, Stip E, et al. Neuropsychological change in
techniques to study the mechanism of cognitive function early phase schizophrenia during 12 months of treatment with
improvements in the brain, which could be a focus of olanzapine, risperidone, or haloperidol. The Canadian
future study. Collaborative Group for research in schizophrenia. Arch Gen
Psychiatry. 2000;57(3):249–258.
In summary, atypical antipsychotics improved certain cog-
[10] Ferreira CD, de Souza MGD, Fernandez-Calvo B, et al.
nitive domains in first-episode schizophrenia, but no signifi- Neurocognitive functions in schizophrenia: a systematic review of
cant difference in the degree of improvements was found the effects of typical and atypical antipsychotic drugs. Psychol
between drugs. Improvements of some cognitive domains Neurosci. 2016;9(1):12–31.
are associated with reduction of negative symptoms, general [11] Han X, Yuan YB, Yu X, et al. The Chinese First-Episode
Schizophrenia Trial: background and study design. East Asian
psychopathological symptoms and improvement of overall
Arch Psychiatry. 2014;24(4):169–173.
clinical symptoms. [12] Kay SR, Opler LA, Lindenmayer JP. Reliability and validity of the
positive and negative syndrome scale for schizophrenics.
Psychiatry Res. 1988;23(1):99–110.
Acknowledgements [13] Lingjaerde O, Ahlfors UG, Bech P, et al. The UKU side effect rating
scale. A new comprehensive rating scale for psychotropic drugs
We thank all the staff in six study sites that were involved in organizing
and a cross-sectional study of side effects in neuroleptic-treated
and recruiting samples as well as all the subjects providing the neurobe-
patients. Acta Psychiatr Scand Suppl. 1987;334:1–100.
havioral data to our study.
[14] Kern RS, Nuechterlein KH, Green MF, et al. The MATRICS
Consensus Cognitive Battery, part 2: co-norming and standardiza-
tion. Am J Psychiatry. 2008;165(2):214–220.
Disclosure statement [15] Shi C, Kang L, Yao S, et al. The MATRICS Consensus Cognitive
The authors report no conflicts of interest. The authors alone are respon- Battery (MCCB): co-norming and standardization in China.
sible for the content and writing of the paper. Schizophr Res. 2015;169(1–3):109–115.
Notes on contributors [16] Wang J, Hu M, Guo X, et al. Cognitive effects of atypical anti-
Yanyan Hou performed the statistical analysis, and drafted the manu- psychotic drugs in first-episode drug-naïve schizophrenic
script. Jiaheng Xie performed the statistical analyses. Yanbo Yuan, Zhang patients. Neural Regen Res. 2013;8(3):277–286.
Cheng, Xue Han and Lei Yang were principal investigators for recruit- [17] Harvey PD, Green MF, McGurk SR, et al. Changes in cognitive
ment. Xin Yu and Chuan Shi provided critical contribution and review of functioning with risperidone and olanzapine treatment: a large-
the manuscript and Xin Yu conceived and designed the trial. All authors scale, double-blind, randomized study. Psychopharmacology
read and approved the manuscript (Berl). 2003;169(3–4):404–411.
[18] Riedel M, Spellmann I, Schennach-Wolff R, et al. Effect of aripipra-
zole on cognition in the treatment of patients with schizophre-
Funding nia. Pharmacopsychiatry. 2010;43(2):50–57.
[19] Matsuda Y, Sato S, Iwata K, et al. Effects of risperidone and aripi-
This work was supported by the National Key R&D Program of China prazole on neurocognitive rehabilitation for schizophrenia.
(2018YFC1314200),the Key Program of Beijing Science and Technology Psychiatry Clin Neurosci. 2014;68(6):425–431.
Commission under Grant No. D171100007017002; the National Key [20] Perlick D, Stastny P, Katz I, et al. Memory deficits and anticholin-
Project of Scientific and Technical Supporting Programs from the ergic levels in chronic schizophrenia. Am J Psychiatry. 1986;
Ministry of Science & Technology of China under Grant No. 143(2):230–232.
2007BAI17B04. AQ1v [21] Strauss ME, Reynolds KS, Jayaram G, et al. Effects of anticholiner-
gic medication on memory in schizophrenia. Schizophr Res. 1990;
3(2):127–129.
References [22] Eum S, Hill SK, Rubin LH, et al. Cognitive burden of anticholiner-
gic medications in psychotic disorders. Schizophr Res. 2017;190:
[1] Green MF, Nuechterlein KH, Gold JM, et al. Approaching a con- 129–135.
sensus cognitive battery for clinical trials in schizophrenia: the [23] Vinogradov S, Fisher M, Warm H, et al. The cognitive cost of anti-
NIMH-MATRICS conference to select cognitive domains and test cholinergic burden: decreased response to cognitive training in
criteria. Biol Psychiatry. 2004;56(5):301–307. schizophrenia. Am J Psychiatry. 2009;166(9):1055–1062.
[2] Albus M, Hubmann W, Mohr F, et al. Neurocognitive functioning [24] Ang MS, Abdul Rashid NA, Lam M, et al. The impact of medica-
in patients with first-episode schizophrenia : results of a pro- tion anticholinergic burden on cognitive performance in people
spective 5-year follow-up study . Eur Arch Psychiatry Clin with schizophrenia. J Clin Psychopharmacol. 2017;37(6):651–656.
Neurosci. 2006;256(7):442–451. [25] Keefe RS, Bilder RM, Davis SM, et al. Neurocognitive effects of
[3] Gross G. Novel antischizophrenia treatments. Handb Exp antipsychotic medications in patients with chronic schizophrenia
Pharmacol. 2012;213(5):13–20. in the CATIE Trial. Arch Gen Psychiatry. 2007;64(6):633–647.
[4] Bora E. Neurodevelopmental origin of cognitive impairment in [26] Heckers S. Neuroimaging studies of the hippocampus in schizo-
schizophrenia. Psychol Med. 2015;45(1):1–9. phrenia. Hippocampus. 2001;11(5):520–528.
[5] Bora E, Akdede BB, Alptekin K. The relationship between cognitive [27] Velakoulis D, Wood SJ, Wong MT, et al. Hippocampal and amyg-
impairment in schizophrenia and metabolic syndrome: a system- dala volumes according to psychosis stage and diagnosis: a mag-
atic review and meta-analysis. Psychol Med. 2017;47(6):1030–1040. netic resonance imaging study of chronic schizophrenia, first-
[6] Green MF. What are the functional consequences of neurocognitive episode psychosis, and ultra-high-risk individuals. Arch Gen
deficits in schizophrenia? Am J Psychiatry. 1996;153(3):321–330. Psychiatry. 2006;63(2):139–149.
[7] Jeste SD, Patterson TL, Palmer BW, et al. Cognitive predictors of [28] Antoniades M, Schoeler T, Radua J, et al. Verbal learning and hip-
medication adherence among middle-aged and older outpatients pocampal dysfunction in schizophrenia: a meta-analysis. Neurosci
with schizophrenia. Schizophr Res. 2003;63(1–2):49–58. Biobehav Rev. 2018;86:166–175.
8 Y. HOU ET AL.

[29] Buchy L, Czechowska Y, Chochol C, et al. Toward a model of cog- [35] Huang M, Kwon S, Oyamada Y, et al. Dopamine D3 receptor
nitive insight in first-episode psychosis: verbal memory and hip- antagonism contributes to blonanserin-induced cortical dopa-
pocampal structure. Schizophr Bull. 2010;36(5):1040–1049. mine and acetylcholine efflux and cognitive improvement.
[30] Cuesta MJ, Jalon EG, Campos MS, et al. Cognitive effectiveness of Pharmacol Biochem Behav. 2015;138:49–57.
olanzapine and risperidone in first-episode psychosis. Br J [36] Elbe D. Stahl’s essential psychopharmacology: neuroscientific
Psychiatry. 2009;194(5):439–445. basis and practical applications, third edition. J Can Acad Child
[31] Keefe RS, Sweeney JA, Gu H, et al. Effects of olanzapine, quetia- Adolesc Psychiatry. 2010;19(3):230.
pine, and risperidone on neurocognitive function in early psych- [37] Sydserff S, Sutton EJ, Song D, et al. Selective alpha7 nicotinic
receptor activation by AZD0328 enhances cortical dopamine
osis: a randomized, double-blind 52-week comparison. AJP. 2007;
release and improves learning and attentional processes.
164(7):1061–1071.
Biochem Pharmacol. 2009;78(7):880–888.
[32] Keefe RSE, Davis VG, Harvey PD, et al. Placebo response and prac-
[38] Censits DM, Ragland JD, Gur RC, et al. Neuropsychological evi-
tice effects in schizophrenia cognition trials. JAMA Psychiatry.
dence supporting a neurodevelopmental model of schizophrenia:
2017;74(8):807–814. a longitudinal study. Schizophr Res. 1997;24(3):289–298.
[33] Lesh TA, Niendam TA, Minzenberg MJ, et al. Cognitive control [39] Harvey PD, Keefe RS. Studies of cognitive change in patients with
deficits in schizophrenia: mechanisms and meaning. schizophrenia following novel antipsychotic treatment. Am J
Neuropsychopharmacology. 2011;36(1):316–338. Psychiatry. 2001;158(2):176–184.
[34] Huang M, Kwon S, He W, et al. Neurochemical arguments for the [40] Schuepbach D, Keshavan MS, Kmiec JA, et al. Negative symptom
use of dopamine D4 receptor stimulation to improve cognitive resolution and improvements in specific cognitive deficits after
impairment associated with schizophrenia. Pharmacol Biochem acute treatment in first-episode schizophrenia. Schizophr Res.
Behav. 2017;157:16–23. 2002;53(3):249–261.