The Association Between Psychotropic Medications and Cognitive Functioning in a Real-World Cohort of 869 Individuals with Schizophrenia

Abstract

Background

Cognitive deficits are a central feature of schizophrenia for which there are not any established pharmacological treatments. Antipsychotics are the mainstay of schizophrenia therapy but the effects of these and other psychotropic medications on the cognitive functioning of people schizophrenia have not been extensively studied in routine real-world settings.

Study Design

A total of 869 people with schizophrenia receiving community-based care were assessed on a cognitive battery, the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). The machine-learning tool of partialing-out least absolute shrinkage and selection operator (LASSO) regression were used to examine the independent association between the RBANS total and index scores and receipt of individual psychotropic medications considering relevant demographic, clinical, and environmental covariates. In this cross-sectional study, we also examined the effects of medication dosage and some medication combinations.

Study Results

We found that 4 medications, clozapine, quetiapine, benztropine, and oral haloperidol were each independently associated with significantly reduced cognitive scores compared with people not receiving these medications. Three of the 4 medications, clozapine, haloperidol, and benztropine, showed a significant dose-related relationship with total cognitive score. We also documented further reductions in cognitive functioning in people receiving some pair-wise combinations of these medications. Reduced memory was the domain most associated with these medications, especially among people receiving clozapine.

Conclusions

Prescribers may consider minimizing doses and limiting the administration of combinations of the identified medications. Interventions should be further developed for people with schizophrenia to improve their cognitive functioning and quality of life.

Introduction

Cognitive deficits are a core feature of schizophrenia.¹ As a group, people with schizophrenia perform more poorly than age-matched controls on measures of verbal memory, executive functioning, attention, processing speed, and other cognitive domains.² Cognitive deficits are associated with worse social functioning outcomes in people with schizophrenia and are a major contributing factor to the high burden of the disease.^3–5 Cognitive deficits are also associated with increased rates of premature mortality.⁶

Antipsychotic medications are the mainstay of schizophrenia treatment.⁷ However, these medications are not effective for the treatment of cognitive deficits even when they result in a diminution of psychotic symptoms.⁸ Furthermore, there are no medications with established efficacy to treat this domain.⁹

Other types of medications are also frequently prescribed to address other schizophrenia-related symptoms such as depression, mood instability, and anxiety. These medications include antidepressants, mood stabilizers, and anxiolytics. Medications such as anti-cholinergic agents may also be prescribed to treat the side effects of antipsychotic medications.

Several previous studies have addressed the effects of psychotropic medications administered to individuals with schizophrenia under clinical trial conditions where they are often compared to a placebo.^8,10,11 However, there have been few studies of the relationship between the administration of these medications to individuals with a diagnosis of schizophrenia, schizophreniform disorder, or schizoaffective disorder in routine community settings taking into account relevant covariates such as psychotic symptoms, educational status, exposure to environmental factors, and comorbid medical conditions. Furthermore, there are few studies that have examined the effects of medication dosage or the receipt of multiple medications.

Early studies indicated that individuals with schizophrenia receiving first-generation antipsychotics had cognitive deficits and social disability which persisted despite improvements in psychiatric symptoms.¹² Second-generation antipsychotics such as olanzapine and risperidone were initially thought to pose less of a cognitive burden in schizophrenia, or even to have a pro-cognitive effect,^13,14 but later determined to have similar effects as the first-generation agents.^15,16 The adverse cognitive effects of anti-cholinergic agents, prescribed for antipsychotic medication side effects, are also well established.¹⁷ The effects of mood stabilizing medications and of antidepressant medications on cognition have been studied most extensively in mood disorders with only limited investigation in people with schizophrenia.^18,19

The reasons for the cognitive effects of antipsychotic and other psychotropic medications are not known with certainty but may include anticholinergic side effects and receptor binding profiles with effects on critical brain regions.^20,21 Adverse cognitive effects of medications may also be due to the sedative properties of some compounds and their metabolic side effects.²²

We examined the association between psychotropic medications and cognitive functioning in a cohort of 869 individuals with schizophrenia, schizophreniform disorder, or schizoaffective disorder receiving psychiatric care from community-based psychiatric programs.

Methods

Study Population

Participants were recruited at Sheppard Pratt, a large private, nonprofit, community-based psychiatric health care system in central Maryland, USA. Participants were enrolled between February 1999 and June 2024. The inclusion criteria were: (1) diagnosis of schizophrenia, schizophreniform disorder, or schizoaffective disorder meeting criteria in the Diagnostic and Statistical Manual of Mental Disorder Fourth Edition (DSM-IV, 1994) based on the Structured Clinical Interview for Diagnosis for Axis I disorders (SCID)²³ and medical records (hereafter referred to as schizophrenia); (2) age 18-65 (18-45 for those with schizophreniform disorder); (3) receiving an anti-psychotic medication. Exclusion criteria were: (1) history of intravenous substance abuse; (2) diagnosis of childhood intellectual disability; (3) clinically significant medical disorder that would affect cognitive performance such as epilepsy, HIV, history of encephalitis or serious head trauma, or any other reported neurological disorder of the central nervous system; (4) primary diagnosis of substance abuse or dependence within the past three months

The study was approved by the Institutional Review Boards of Sheppard Pratt and the Johns Hopkins Medical Institutions following established guidelines. All participants provided written informed consent after the study procedures were explained.

Demographic and Clinical Measures

Participants were seen for one study visit at which time they were interviewed and medical charts reviewed; information was obtained about demographic variables, co-occurring medical conditions from a review of systems, current tobacco smoking, and history of substance use. Psychiatric symptoms were assessed with the Positive and Negative Syndrome Scale (PANSS).²⁴ Body mass index (BMI) was calculated from measured weight and height.

Medications

The psychotropic medications and dosage of each psychotropic medication prescribed to the participant at the time of the study visit were obtained from the medical record and participant interviews.

Cognitive Assessment

At the study visit, participants were administered a cognitive test battery, the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS).²⁵ The RBANS was selected for the measurement of cognitive functioning since it is sensitive to the level and type of cognitive impairments which are often found in individuals with schizophrenia.²⁶ Test indexes are Immediate Memory (comprised of List Learning and Story Memory tasks); Visuospatial/Constructional (comprised of Figure Copy and Line Orientation tasks); Language (comprised of Picture Naming and Fluency tasks); Attention (comprised of Digit Span and Coding tasks); and Delayed Memory (comprised of List Recall, Story Recall, Figure Recall, and List Recognition tasks). Each index score is expressed as an age-adjusted standard score with a mean of approximately 100 and a standard deviation of approximately 15 based on a previously described normative study group of 540 healthy subjects, ranging in age from 20 to 89, matched to the US Census on gender, ethnicity, and level of education. The index scores are combined to yield a Total RBANS score which is a measure of overall cognitive functioning. Participants were administered Form A of the RBANS.

Serological Assessment

Participants had a blood sample drawn which was employed to measure antibodies to Herpes Simplex Virus type 1 employing a type-specific enzyme immunoassay. The results were classified as positive or negative based on comparison with a standard sample included in each assay run as previously described.²⁷ This antibody was chosen for measurement based on previous studies documenting its association with cognitive functioning in individuals with schizophrenia.²⁸

Statistical Analysis

We employed the machine-learning tool of partialing-out least absolute shrinkage and selection operator (LASSO) regression to measure the association between cognitive functioning as measured by the RBANS Total Score and the receipt of psychotropic medications along with relevant covariates.²⁹ The medications evaluated included all those received by at least 5% of the population at the time of assessment. This method of statistical analysis was chosen since it allows for the identification of the independent associations of multiple variables of interest and covariates in an efficient manner with a minimum of confounding. Missing data in the covariates were estimated using multiple imputations. The variables included as covariates along with the number of imputed covariates are shown in Table 1.

Table 1.

Demographic, Clinical, and Cognitive Characteristics, and Medications Received by at Least 5% of the Sample (N = 869)¹

Demographic variables
Age at enrollment, years	39.6 ± 12.0
Race
White	489 (56.3%)
Black	350 (40.3%)
Other	30 (3.4%)
Female sex	311 (35.8%)
Education 2
< 12 years	233 (26.8%)
High school graduate only	313 (36.0%)
Some college	219 (25.2%)
College graduate	104 (12.0%)
Maternal education
< 8 years	55 (6.3%)
>=8 years - < 12 years	450 (51.8%)
>=12 years - < 16 years	313 (35.0%)
>= 16 years	41 (5.9%)
Schizophrenia diagnosis 3
Schizophreniform disorder	32 (3.7%)
Schizoaffective disorder	394 (45.3%)
Schizophrenia	443 (51.0%)
Clinical variables
Care setting
Outpatient	655 (75.3%
Day hospital	57 (6.6%)
Inpatient unit	157 (18.1%)
Age of illness onset 4	20.9 ± 7.6
Duration of illness, years	18.5 ± 11.8
PANSS 5 Positive symptom score	19.0 ± 5.2
PANSS 5 Negative symptom score	20.2 ± 4.9
PANSS 5 General symptom score	35.6 ± 7.7
PANSS 5 Total symptom score	74.9 ± 14.2
Substance use history 6
1 No history of substance misuse	309 (35.6%)
2 Substance misuse but not in past 3 mos.	451 (51.9%)
3 Substance misuse in past 3 mos.	109 (12.5%)
Tobacco smoker	535 (61.6%)
Body mass index 7	30.7 ± 6.2
Seropositivity to Herpes Simplex Virus type 1	368 (42.4%)
Comorbid medical conditions
Musculoskeletal	119 (13.7%)
Endocrine	375 (43.0%)
Genitourinary	87 (10.0%)
Respiratory	181 (20.8%)
Neoplastic	23 (2.7%)
Cardiovascular	401 (46.1%)
Gastrointestinal	311 (35.8%)
Hematologic	56 (6.4%)
Immune	58 (6.7%)
Hepatic	37 (4.3%)
Neurological	134 (15.4%)
Dermatologic	96 (11.1%)
Cognitive assessment
RBANS 8 Total	65.6 ± 13.0
RBANS 8 immediate memory	67.0 ± 16.8
RBANS 8 visuospatial/constructional	70.4 ± 14.4
RBANS 8 attention	72.6 ± 16.4
RBANS 8 language	81.8 ± 15.5
RBANS 8 delayed memory	69.0 ± 18.2
Psychotropic medications 9
Antipsychotic medications
Clozapine	173 (19.9%)
Olanzapine—oral	205 (23.6%)
Quetiapine	108 (12.4%)
Risperidone—oral	234 (26.9%)
Aripiprazole—oral	66 (7.6%)
Fluphenazine—oral	65 (7.5%)
Haloperidol—oral	104 (12.0%)
Antidepressant medications
Fluoxetine	52 (6.0%)
Paroxetine	47 (5.4%)
Sertraline	64 (7.4%
Trazadone	81 (9.3%)
Mood stabilizing medications
Lithium	112 (12.9%)
Valproic acid	185 (21.3%)
Antiparkinsonian medication
Benztropine	277 (31.9%)
Anxiolytic medications
Clonazepam	96 (11.0%)
Lorazepam	59 (6.8%)

¹Mean ± s.d. or number (%); all variables listed in the table were included as covariates except for the RBANS variables which were the outcomes.

²1 missing, maternal education used.

³DSM IV criteria.

⁴6 values imputed.

⁵PANSS, Positive and Negative Syndrome Scale.

⁶1 missing, rating of 1 used.

⁷290 values imputed.

⁸RBANS, Repeatable Battery for the Assessment of Neuropsychological Status.

⁹PRNs not included.

Medications showing a statistically significant association (alpha = .05, adjusted for multiple comparisons) with RBANS Total score were further analyzed. We employed similar LASSO models to examine the relationship between RBANS Total score and the receipt of pairs of these medications. We also used LASSO models to examine the relationship between the dosage of these medications and the RBANS Total score employing 3 dosage groupings based on the distribution of dosages in the study population as shown in Table S3. We employed similar LASSO models to examine the relationship between these medications and the RBANS index scores. The initial LASSO regression model examining RBANS Total score included all of the 29 demographic, clinical, and medical comorbidity variables as well as all of the 16 medications listed in Table 1. Follow-up LASSO regression models also included the same covariates and medications. The composition of each of the Lasso regression models is described in the supplementary material.

The LASSO regression models were used to assign regression coefficients, 95% confidence intervals, and probability estimates (P values).³⁰ Wald tests were further employed to determine the statistical significance of dose-related effects and the relationship among the different medication combinations as described in Table S1. All statistical procedures were performed using the “ivporegress” and “test” algorithms available in STATA version 18. Significance values (P values) generated by the LASSO models were adjusted for multiple comparisons using the false discovery method of Benjamini et al.³¹ An adjusted P value of ≤.05 was considered to represent statistical significance. Since the RBANS Total score and index scores are based on a mean level of 100, the regression coefficients can be interpreted as an estimate of the percentage change in the RBANS scores independently associated with the individual medication.

Results

A total of 869 participants were enrolled in the study; their demographic and clinical characteristics and the psychotropic medications that were received by at least 5% of the study population are shown in Table 1. A list of the psychotropic medications received by less than 5% of the participants is found in S2. The total number of psychotropic medications received by study participants is presented in Table S3.

The initial LASSO regression model examined the relationship between RBANS Total score and the covariates and medications listed in Table 1. As depicted in Figure 1, this model identified significant independent associations between a lower RBANS Total score and the receipt of 4 of the psychotropic medications analyzed. These medications were: clozapine (coefficient −4.51, 95% CI −6.53, −2.51, adjusted P = .001); quetiapine (coefficient −2.96, 95% CI −5.17, −0.75, adjusted P = .0144); haloperidol (coefficient −2.72, 95% CI −4.88, −.56, adjusted P = .048); and benztropine (coefficient −4.59, 95% CI −6.17, −3.01, adjusted P = .001). None of the other antipsychotic medications showed a significant association with RBANS Total score. There were no significant associations between RBANS total score and any of the mood stabilizing, antidepressant, or anxiolytic medications. The LASSO regression model also identified race/ethnicity, decreased levels of maternal and participant education, and increased PANSS negative symptom scores and HSV-1 seropositivity as significant covariates contributing to the RBANS Total score.

Figure 1.

Association between Psychotropic Medications Received by at Least 5% of the Study Population and RBANS Total Score. Means, 95% confidence intervals, and significance levels were determined by a LASSO regression model as defined in the text: *** adjusted P < .001, * adjusted P < .05.

We further employed LASSO regression models to examine the relationship between the dosage range of these 4 medications and RBANS Total score as shown in Figure 2. Clozapine, haloperidol, and benztropine all showed significant associations between dose range and RBANS total score with the highest dose range associated with the lowest score. The dose-related association of quetiapine and RBANS total score did not reach statistical significance.

Figure 2.

Association Between Medication Dosage Ranges and RBANS Total Score. The relationships were calculated by LASSO regression models for the 4 medications which showed association with RBANS Total score as described in the text. Dose ranges for each medication are based on the distribution of doses within the study sample. The dosage ranges analyzed and the number of individuals in each dosage range are were as follows: Clozapine Low = < 250 (n = 50), Middle = < 250 - <= 500, (n = 86) High = > 500 - <=750 (n = 43); Quetiapine Low = <=250 (n = 30), Middle = < 250 - <=500 (n  = 32), High = > 500-<=750 (n  = 43); Haloperidol Oral Low = <=5 (n  = 18), Middle = > 5 - < 1 (n  = 35), High = > 10 (n  = 48); Benztropine Low = <=1 (n  = 85), Middle = > 1 - <=2 (n  = 105), High = > 2 (n  = 53). Statistical significance for differences among the doses for each medication employing the Wald test; ***P < .001, **P < .01.

We also examined the relationship between the receipt of pairs of these medications at any dose and the RBANS total score. As depicted in Figure 3, two of these combinations, benzotropine + haloperidol (P < .001) and benzotropine + quietapine (P < .005) demonstrated a statistically lower RBANS total score that was greater than that associated with the receipt of either of the individual medications comprising the combination. The effects of the other 2-way combinations were not significantly greater than those of the individual medications. We did not have a sufficient sample size to calculate the effect of the receipt of three-way or these medications or all 4 medications.

Figure 3.

Association Between Individual Medications and Medication Combinations and RBANS Total Score. The LASSO regression coefficients of the RBANS Total Scores for the combinations of medications which showed significant individual associations with the RBANS Total Scores are depicted along with the coefficients of the individual medications. The combinations and the number of individuals with each combination are as follows: clozapine + benztropine (n  = 27), clozapine + haloperidol (n  = 20), haloperidol + benztropine (n  = 76), quetiapine + benztropine (n  = 34), clozapine + quetiapine (n  = 12), haloperidol + quetiapine (n  = 12). * Indicates P < .05 of the combination compared to each of its individual components as calculated by Wald tests.

We also employed similar LASSO regression models to examine the RBANS index scores associated with the receipt of the 4 medications which were associated with lower RBANS Total scores. The RBANS index scores measure cognitive domains of Immediate Memory, Delayed Memory, Attention, Language, and Visuospatial/Constructional. As shown in Figure 4, each of the 4 medications was associated with Immediate Memory and all except quetiapine were associated with Delayed Memory. However, there were differences among the medications in terms of the relationship to other index scores. Receipt of clozapine was associated with lower scores on the Immediate Memory (coefficient −10.17, 95% CI −12.69, −7.64, adjusted P < .003) and the Delayed Memory indexes (coefficient −6.03, 95% CI −9.38, −3.23, adjusted P < .003), but not with the other index scores. Receipt of quetiapine was associated with lower scores on the Immediate Memory index (coefficient −3.85, 95% CI −6.97, −0.73, adjusted P = .01) and the Language index (coefficient −3.97, 95% CI −6.93, −1.01, adjusted P = .029) but was not associated with Delayed Memory or any of the other index scores. Receipt of haloperidol was associated with lower Immediate Memory (coefficient −3.52, 95% CI −6.45, −0,58, adjusted P < .01) and Delayed Memory indexes (coefficient −4.58, 95% CI −8.25, −0.92, adjusted P = .009). Benztropine was significantly associated with all of the index scores except Language: Immediate Memory (coefficient −6.85, 95% CI −9.05, −4.64, adjusted P < .003), Delayed Memory (coefficient −4.89, 95% CI −7.45, −2.34, adjusted P < .003), Visuospatial/ Constructional (coefficient −3.51, 95% CI −5.26, −1.77, P < .005); and Attention (coefficient −3.82, 95% CI −6.05, −1.58, adjusted P = .005). The covariates identified in these models were similar to those found for the RBANS Total score and are shown in Table S4. The covariates identified in the analyses of the individual medications and Total cognitive score are shown in Table S5.

Figure 4.

Association Between Individual Medications and RBANS Index Scores. Means, 95% confidence intervals, and significance levels for the indicated Index scores were determined by LASSO regression models as defined in the text, *** adjusted P < .001, ** adjusted P < .01, * adjusted P < .05.

Discussion

We employed the machine learning tool of partialing-out LASSO regression to examine the independent association between psychotropic medications and cognitive functioning in a cohort of 869 people with schizophrenia enrolled from community-based psychiatric settings. This method combines LASSO to identify relevant covariates followed by partialing-out regression to determine coefficients and statistical significance.^29,30 LASSO-based methods were selected for this study since they overcome some limitations inherent in the use of linear regression models which contain large numbers of variables of interest. These limitations include inadequate sample size, overfitting, multiple collinearity, and model complexity.^32,33 This model allowed us to incorporate the effects of several clinical, demographic, and environmental factors that can affect cognitive functioning.

We found that 4 medications, clozapine, quetiapine, haloperidol, and benztropine, were each independently associated with reduced RBANS Total cognitive scores. In terms of cognitive domains. reduced Immediate and Delayed Memory were the RBANS index scores most associated with these medications. Three of the medications, clozapine, haloperidol, and benztropine, showed a significant dose-related relationship with Total cognitive score. We also found further reduced RBANS Total cognitive scores in people receiving combinations of benzotropine and haloperidol as well as benzotropine and quietapine in which cases the effect of receiving the combination was greater than that of receiving the individual medication.

Clozapine is a second-generation antipsychotic medication used for people with treatment resistant schizophrenia.³⁴ In our sample, receipt of clozapine was significantly and inversely associated with an approximately 5% decrease in RBANS Total score. Clozapine was also associated with significantly lower Immediate and Delayed Memory index scores (by approximately 10% and 6%, respectively) but not the other L index scores. Clozapine has been shown to be a uniquely effective antipsychotic agent, but its effects on cognitive functioning have not been widely investigated. Early studies with relatively small samples and short-term follow-up periods suggested that clozapine improved cognitive performance in the domains of attention and verbal fluency.^13,35 However, in an open label follow-up study, these gains were not sustained over time, and short-term memory declined over a mean follow-up interval of 14 years.³⁶ Also, two recent meta-analyses of randomized controlled trials comparing clozapine to another antipsychotic or against a placebo found that people receiving clozapine had lower cognitive performance than people in the comparison conditions.^10,11 The greater illness severity of clozapine recipients than other patients with schizophrenia has been suggested as an explanation for their poorer cognitive functioning³⁷; however, in this study, symptom severity, duration of illness, and other clinical factors were taken into account in the analysis. We cannot exclude other factors associated with illness severity which have impacted clozapine usage.

Consistent with the results of our study, other cross-sectional studies have shown that clozapine users have poorer short-term verbal memory and delayed memory than people receiving other antipsychotic medications^38,39 In our study, the effects on memory associated with clozapine were greater than on the other cognitive domains which were relatively spared (see Figure 4).

We also found an association between receipt of quetiapine and lower Total cognitive score (by approximately 3%) and Immediate Memory and Language Index scores (both by approximately 4%; Figure 4). Previous studies comparing the cognitive effects of quetiapine to those of another individual antipsychotic or to placebo have had mixed results. One recent meta-analysis found that persons with schizophrenia receiving quetiapine performed better on a composite cognitive measure than people receiving other antipsychotics¹¹; however, a more recent meta-analysis did not show this result and the performance of persons receiving quetiapine was mid-range on cognitive outcomes among other antipsychotic medications.¹⁰ The trials included in these meta-analyses were randomized controlled trials including registration trials and examined changes in cognitive score over relatively short time periods.

Our finding regarding the significant negative association between the receipt of haloperidol and cognitive performance is consistent with other trials in showing the adverse cognitive effects of first-generation antipsychotic medications for people with schizophrenia, especially haloperidol as it has been the first-generation antipsychotic medication used most frequently in comparison trials.^11,40 In our study, we found a negative association not only with Total cognitive score (approximately 5% lower score) but also with Immediate Memory, Delayed Memory, Language, and Visuospatial/Constructional Index scores (lower by approximately 7%, 5%, 4%, and 4%, respectively). To the best of our knowledge, there are few recent studies comparing the performance of people receiving oral preparations of haloperidol to those receiving other antipsychotics in routine, real-world, community settings. Our results here are consistent with those found in head-to-head randomized trials in which the cognitive performance of people receiving haloperidol was found to be consistently below that of people receiving other individual antipsychotic medications.^11,41

Benztropine, the only antiparkinsonian medication received by at least 5% of people in our sample, was independently associated with lower cognitive score (by approximately 6-7%). Antiparkinsonian agents including benztropine are frequently prescribed to treat and prevent anti-psychotic induced extrapyramidal symptoms. While effective for the treatment and prevention of these side effects, these and other anti-parkinsonian medications have been associated with lower cognitive performance in patients with schizophrenia, consistent with the results of our study.^17,19,42 In addition, several small studies indicate that discontinuing antiparkinsonian agents such as benztropine in people with schizophrenia can have a beneficial effect on cognitive functioning.^43,44 Also consistent with previous studies is the pervasive effect of benztropine on cognitive functioning; all of the RBANS index scores except the Language index score were significantly lower in people receiving benztropine (Figure 4). In addition, benztropine was a medication in three medication combinations (Figure 3) which showed a more negative association with cognitive functioning than either medication alone, suggesting that it adds to the adverse cognitive effects of other medications.

In most placebo-controlled trials, patients are treated with the optimum dosage, precluding an investigation of dose-related effects on cognition. We found a dose-related effect on Total cognitive score for clozapine, haloperidol, and benztropine; the dose-related effect of quetiapine showed a trend towards more impairment at higher doses but did not reach statistical significance (Figure 2). To the best of our knowledge, there are few previous studies examining the dose-related effects of psychotropic medications on cognition in schizophrenia in routine community settings. It is possible that the higher doses were indicative of more severe psychiatric symptoms leading to the prescription of increased doses, confounding by indication. However, the dose-related effects of these medications were found to be independent of symptom scores suggesting that increased current symptom severity cannot totally explain the dose-related association of the medications with cognitive functioning. Our results are also consistent with studies showing an adverse cognitive effect of lifetime antipsychotic dose-years; follow-ups of people with schizophrenia in the 1966 Northern Finland Birth Cohort found that after an average of 9 years and of 16 years of illness duration, higher antipsychotic dose-years were associated with significantly worse cognitive scores.^45,46 However, these studies did not look at the prescribed doses, the effects of individual antipsychotic medications, or the effect of other psychotropic agents.

An explanation for the adverse effect of these medications on cognitive functioning is not known with certainty but may be related to the receptor-binding profiles of these medications and their combined effects on critical neurotransmitter systems and brain regions. These medications disrupt the balance of multiple neurotransmitter pathways, including dopaminergic, glutamatergic, and cholinergic systems, all of which are integral to cognitive function. Clozapine, for instance, is associated with pronounced deficits in immediate and delayed memory, likely due to its anticholinergic properties, given the essential role of acetylcholine in encoding and retrieval processes.⁴⁷ Similarly, benztropine’s strong anticholinergic effects disrupt acetylcholine signaling, resulting in widespread cognitive impairments¹⁷ Haloperidol’s cognitive deficits are likely driven by its potent dopamine blockade in the prefrontal cortex, a region critical for working memory and executive function.⁴⁸ In contrast, quetiapine exhibits relatively modest impairments in immediate memory and language, with no clear dose-response relationship. This pattern may reflect its lower antimuscarinic activity compared to clozapine and reduced dopamine receptor affinity relative to haloperidol.⁴⁹ More extensive consideration of these issues is found in other research studies.

Strengths of our study include a relatively large sample of people with schizophrenia, comprehensive and thorough collection of medication data, a standardized protocol and data collection procedures, and standardized and consistent RBANS scoring throughout the study period. In addition, we applied a machine learning data analytic method that allowed for the inclusion of many covariates and controlled for the receipt of multiple medications. In addition, we were able to analyze the effects of several medication dose ranges and medication combinations. We also had few exclusions to study participation, so our population was likely representative of people with schizophrenia receiving care in routine, community e settings, albeit not including those with recent alcohol or other substance misuse.

Limitations of our study include that our cognitive battery was relatively brief and did not include a measure of executive functioning. In addition, the medications analyzed were only those that were received by 5% of the study population, so the possible effects of some of the less frequently used antipsychotics were not included in the analysis. Sample size may also have limited the ability to detect significant associations between cognitive functioning and the other medications included in the study as well as the effects of other medication combinations. We also did not examine long-acting parenteral medications, none of which were received by 5% of our sample. We also did not have a measure of medication adherence or medication blood levels, However, participants were receiving psychiatric care, and the significant dose-related effects give credence to the findings. In addition, due to the cross-sectional nature of our study, we could not examine the effects of medications on changes in cognitive functioning over time, nor did we include measures of illness severity such as social functioning or quality of life Furthermore, while we controlled for a large number of variables, it is possible that unmeasured variables could contribute to the associations with cognitive functioning. For example, we did not measure the total anticholinergic burden of the medications received or patient effort on cognitive tests which can affect cognitive test performance^17,50 Finally, the medications were not prescribed following a research protocol but were prescribed by clinicians in community programs based on their clinical judgment. However, we believe that this also is a strength of our study in that it reflects the experience of people with schizophrenia in real-world settings.

Conclusions

We identified 4 psychotropic medications, clozapine, quetiapine, haloperidol, and benztropine, that individually and in combination are associated with poorer cognitive functioning in people with schizophrenia in community-based care. While these medications clearly have their benefits, prescribers may consider minimizing doses and limiting the administration of these medications in combination particularly when the combinations include benztropine. Additional interventions, both pharmacologic and psychosocial, should be further developed and evaluated for people with schizophrenia to improve their cognitive functioning and their quality of life.

Supplementary material

Supplementary material is available at https://academic.oup.com/schizophreniabulletin.

Funding

This study was funded by the Stanley Medical Research Institue (grant # 07-1690 to F Dickerson).