Immune Cell Alterations in Psychotic Disorders: A Comprehensive Systematic Review and Meta-Analysis

BACKGROUND: A comprehensive meta-analysis on the composition of circulating immune cells from both the myeloid and the lymphoid lines including specialized subsets in blood and cerebrospinal ﬂ uid (CSF) of patients with psychotic disorders compared with healthy control participants has been lacking. METHODS: Multiple databases (PubMed, EMBASE, Cochrane Library, Web of Science, ClinicalTrials.gov, and PsycINFO) were searched for eligible studies up until October 18, 2022. All studies investigating circulating immune cells in the blood and CSF from patients with psychotic disorders (ICD-10: F20 and F22-29) compared with healthy control participants were included. RESULTS: A total of 86 studies were included in the meta-analysis. In the blood, the following categories of immune cells were elevated: leukocyte count (31 studies, standardized mean difference [SMD] = 0.35; 95% CI, 0.24 to 0.46), granulocyte count (4 studies, SMD = 0.57; 95% CI, 0.12 to 1.01), neutrophil granulocyte count (21 studies, SMD = 0.32; 95% CI, 0.11 to 0.54), monocyte count (23 studies, SMD = 0.40; 95% CI, 0.23 to 0.56), and B lymphocyte count (10 studies, SMD = 0.26; 95% CI, 0.04 to 0.48). Additionally, the neutrophil/lymphocyte ratio (23 studies, SMD = 0.40; 95% CI, 0.19 to 0.60), the monocyte/lymphocyte ratio (9 studies, SMD = 0.31; 95% CI, 0.04 to 0.57), and the platelet/lymphocyte ratio (10 studies, SMD = 0.23; 95% CI, 0.03 to 0.43) were elevated. The CSF cell count showed a similar tendency but was not signi ﬁ cantly elevated

https://doi.org/10.1016/j.biopsych.2023.11.029Schizophrenia has an estimated worldwide prevalence of around 1% (1).Despite psychotic disorders having a severe impact on the lives of those who are affected, the etiology and pathophysiology behind the disorder remain largely unknown (2,3).It has long been hypothesized that an inflammatory pathophysiological mechanism may be involved in schizophrenia (4,5).Autoimmune diseases and severe infections have been shown to increase the subsequent risk of developing psychosis (6,7).Systematic reviews and meta-analyses have consistently found increased levels of peripheral proinflammatory cytokines such as interleukin 6 in patients with psychotic disorders compared with healthy control participants, especially in patients with first-episode psychosis and acute exacerbation of psychosis (8,9).Anti-inflammatory agents as an add-on treatment to antipsychotic medication have also been shown to have a beneficial effect on psychotic symptoms (10), although the effect was mostly detected in smaller studies.Genetic studies have also found associations between mutations in genes coding for CD19 1 and CD20 1 B lymphocytes and schizophrenia (11).
Four smaller systematic reviews and meta-analyses have been published on circulating immune cells in patients with psychosis compared with healthy controls (12)(13)(14)(15).Two of these articles investigated ratios of cells rather than counts (14,15).Mazza et al. (13) mainly investigated the monocyte count in patients with psychosis.Jackson et al. (12) investigated counts of the major leukocyte subsets (monocytes, neutrophils, eosinophils, and lymphocytes) but did not investigate differences in subsets of lymphocytes that can be differentiated by flow cytometry (i.e., B cells, CD4 1 helper T cells, CD8 1 cytotoxic T cells).A decade-old study by Miller et al. (16) is the only meta-analysis to have investigated potential differences in subsets of lymphocytes, but it included only 16 studies, and many studies on the subject have been published since.New parameters including the Systemic Immune-Inflammation Index have been applied extensively to cancer and other disease entities, while it has been investigated to a lesser degree in psychiatric disorders, and no metaanalyses have reported on it.Moreover, there is growing evidence of greater permeability of the blood-brain barrier in individuals with psychosis (17,18); however, no previous systematic reviews and meta-analyses have investigated the amount of immune cells in the cerebrospinal fluid (CSF).
Therefore, the objective of this first comprehensive systematic review and meta-analysis was to investigate potential differences in the composition of immune cells including subsets measured in the circulating blood and CSF of patients with psychotic disorders compared with healthy control participants.We also investigated whether the number of immune cells was associated with symptom severity in psychosis.

METHODS AND MATERIALS
The study protocol was published a priori on PROSPERO (ID: CRD42023407988).The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and MOOSE (Meta-Analysis of Observational Studies in Epidemiology) guidelines were followed in the writing of this article (Tables S1 and S2).

Study Selection and Search Method
Clinical studies were included based on the following criteria: 1) contained an investigation of circulating immune cells in blood and/or CSF from live human study participants; 2) included patients with a primary diagnosis of psychosis according to diagnostic systems used in the study (ICD, DSM, Research Diagnostic Criteria, or similar classifications) of both sexes and all races; 3) included patients age 18 years or older; 4) included comparison to a healthy control group; 5) published in a peer-reviewed journal; and 6) published in English.
Studies were excluded if they met any of the following criteria: 1) contained data published previously.If studies had overlapping study samples, then the study with the smaller study sample was excluded; 2) investigated psychosis in participants with an affective disorder; 3) investigated psychosis in participants with severe medical conditions, for example, autoimmune, malignant, infectious, neurological, vascular, or endocrine diseases that could have an impact on the circulating immune cells; 4) investigated pregnant or lactating women; 5) had a control group that was admitted at a hospital; and 6) used methods in which the composition of immune cells was not directly measured.
The search was conducted on PubMed, Web of Science, EMBASE, PsycINFO, Cochrane Library, and ClinicalTrials.gov.Results were not restricted based on publication date.The search was concluded on October 18, 2022.The search was conducted using the following medical subject headings Studies conducted on animals, case studies, review articles, and conference abstracts were excluded.An information specialist from the Library of Copenhagen University was consulted in the creation of the search string.The reference lists of relevant articles were searched for additional articles.Articles were screened based on title and abstract by 1 author (MC), and 2 authors (MC and MdR) reviewed the full-text reports to determine whether the studies met inclusion and exclusion criteria.In cases of disagreement on study inclusion, a third author (MEB) was consulted to reach a consensus.

Objectives
The primary objective of the study was to investigate the number of circulating immune cells in the blood and/or CSF of patients with psychosis in comparison to healthy control participants.A secondary objective was to investigate the potential association between immune cell count and the severity of psychosis.

Data Extraction and Bias Assessment
Data extraction was performed by one author (MC), while another author (MdR) inspected and validated all extracted data, which included characteristics of the participants and bibliographical data.The authors were contacted by e-mail in the case of missing details.If no response was received, a reminder was sent.The risk of bias was assessed according to the Newcastle-Ottawa Quality Assessment Scale for casecontrol studies (19) by 2 authors (MC and MdR).Funnel plots for cell counts, percentages, and ratios were created to assess the risk of publication bias when 10 or more studies were included.Subgroup analyses were conducted comparing studies with a Newcastle-Ottawa scale score $ 4 to studies with a score , 4 and studies published before and after the year 2000.Meta-regression was performed to assess the association between the size of the studies and publication year with the standardized mean difference (SMD) for lymphocyte, T lymphocyte, helper T lymphocyte, cytotoxic T lymphocyte, B lymphocyte, leukocyte, neutrophil, and monocyte counts.

Statistical Analysis
SMDs were used to perform meta-analysis using randomeffects models.Results using the random-effects method are presented in the main text and figures because this method includes the between-study variance in the computation of the confidence interval of the overall effect if there is heterogeneity among the studies while the fixed-effect method does not.The results of the fixed-effects method are included in the Supplement for completeness.The variance between studies, s, was calculated using the restricted maximum-likelihood estimator (20), and studies were combined using the inversevariance weighting method.SMD estimates were bias corrected using Hedges' g (21,22).c 2 tests for heterogeneity were performed, and heterogeneity was assessed by calculating I 2 (23).Funnel plots were used to assess possible publication bias.Medians, quartiles, and ranges were used for imputation of means and standard deviations when necessary.Subgroup analyses of inpatient versus outpatient status was performed to investigate associations between psychosis severity and immune cell counts, percentages, and ratios if this information was Meta-Analyses of Immune Cells in Psychotic Disorders available in 10 or more studies.All statistical analyses were conducted in R version 4.1.2using the meta-package version 6.0 (24).All p values were 2-sided, and p values , .05 were considered significant.

RESULTS
A total of 16,585 articles were identified.After the removal of duplicates and screening by title and abstract, 467 studies A Immune cell counts in the blood investigated in 2 studies  were included for full-text inspection, 86 of which could be included in the meta-analysis (Figure S1).Additional information on the individual studies including study and patient characteristics can be found in Table S3, including a list of the markers used to identify the immune cells in the included studies (Tables S9-12).The 86 studies included 25,674 patients overall with a mean age of 34.9 years, while 17,217 healthy control participants were included with a mean age of 34.5 years.Forest plots illustrating the significant results regarding absolute count of immune cells investigated in 2 or more studies based on random-effects models are presented in Figures 1-5.Forest plots for immune cell counts, percentages, and ratios with results from both the fixedand random-effects models can be found in Figures S3-S8.An overview of the results for the immune cell counts, percentages, and ratios can be found in Table 1 and in Tables S4, S5, and S6, respectively.Results from the sensitivity analyses can be found in Figures S58-S139, while results from meta-regression analyses can be found in Figures S140-S177.

Cells in the CSF
Counts of cells in the CSF of patients with psychosis were not significantly different from those of healthy control participants (3 studies, SMD = 0.14; 95% CI, 20.04 to 0.32; I 2 = 0%) (Figure S14).Subsets of cells were investigated in only 1 study, so meta-analyses were not possible.

Sensitivity Analysis on Severity of Psychosis
Patient setting was used as a proxy marker for the severity of the psychosis.Only studies for which the setting was clearly stated were included in the analysis, and of these, few studies included outpatients exclusively (6 studies).The subgroup

Random effects model
Heterogeneity:

Meta-Analyses of Immune Cells in Psychotic Disorders
Biological Psychiatry analysis of inpatient versus outpatient versus mixed showed that the lymphocyte count was significantly increased in the mixed group, with a significant subgroup difference (p = .007)(Figure S93), while there was no difference between inpatients and outpatients exclusively.The T lymphocyte percentage was not significant for inpatients, while it was increased in the mixed group and decreased in outpatients compared with healthy control participants, with a significant subgroup difference (p , .001) (Figure S95).No other significant subgroup differences regarding severity were found for the immune cell counts (Figures S93-S100).Not enough studies reported mean Positive and Negative Syndrome Scale scores to allow us to perform meta-regression.Subgroup analyses of the percentages of patients who were receiving antipsychotic medication in studies did not show any significant differences (Figures S133-S139).S132).The analysis showed a significant subgroup difference for the CD4 1 /CD8 1 ratio (p = .01)(Figure S123) and the helper T lymphocyte count (p = .01)(Figure S130).A subgroup analysis of studies published before and after the year 2000 showed significant subgroup differences for the T lymphocyte count (p = .02)(Figure S84) and the helper T lymphocyte count (p = .006)(Figure S90).Meta-regression on the effect size and the publication year showed significant associations between helper T lymphocyte count and publication year (p = .0017)(Figure S177).Meta-regression showed no association between the effect size and the size of the studies (Figures S140-S158).
The leave-one-out sensitivity analysis altered the results significantly for the platelet/lymphocyte ratio, B lymphocyte count, and helper T lymphocyte count (Figures S71, S72, and S74).

DISCUSSION
To our knowledge, this is the largest and most comprehensive meta-analysis reported to date on the composition of immune cells in patients with psychosis, comprising a total of 86 studies.The results indicate a broad activation of the immune system in patients with psychosis, with the differences in the number of cells not being restricted to a specific line or subset of immune cells and with cells from both the myeloid and the lymphoid line being significantly elevated.The results show elevated levels of leukocytes, granulocytes, neutrophil granulocytes, monocytes, and B lymphocytes, while only the lymphocyte percentage was significantly reduced.Overall, several immune cell alterations were observed, although they may only be present in a subgroup of patients with psychosis.
The previous largest meta-analysis ( 12) consisted of results from 24 studies, whereas we were able to include results from 86 studies.Thus, we are the first to investigate more specialized immune cell subsets like the Th1, Th2, Th17, and regulatory T lymphocytes, with the percentage of both the Th17 and regulatory T lymphocytes being significantly elevated in the patient group.Interestingly, we found the count of the B lymphocytes to be significantly elevated, which is supported by a genome-wide association study that found associations between genes coding for the B lymphocytes (CD19 and CD20) and schizophrenia (11).A previous study that investigated the potential effect of add-on treatment with the COX-2 inhibitor celecoxib also found a significant positive correlation between the decrease in CD19 1 cells and the decrease in negative symptoms as measured by the Positive and Negative Syndrome Scale score (25).Additionally, circulating immune cells like the CD8 1 and CD17 1 T lymphocytes have been shown to be significantly increased in meta-analyses in known neuroinflammatory disorders such as multiple sclerosis (26,27).Furthermore, the scope of our study enabled us to investigate the number of white cells in the CSF, which is of particular interest because of its closer proximity to the brain.
Changes in the immune cells of peripheral blood may affect the cell count in CSF, and while the count of cells in the CSF showed a tendency toward being elevated, the difference was not significant.However, it should be noted that the composition of immune cells in the CSF of patients with schizophrenia have been investigated to a lesser degree than in the blood and that our systematic search for articles led us to only 3 studies that included healthy control groups.Many studies have not included a healthy control group because of the ethical considerations involved in performing a lumbar puncture on healthy volunteers instead of opting to include patients with noninflammatory afflictions as control groups instead (28)(29)(30)(31)(32)(33)(34)(35)(36)(37).In accordance with our results, one of these studies found that monocytes were significantly elevated in the peripheral blood of patients with schizophrenia spectrum disorders, and this was associated with a higher number of monocytes in the CSF (35).These findings were used to make a machine learning model that could distinguish between the schizophrenia spectrum disorders and the control group of patients with idiopathic intracranial hypertension, which opens the future possibility of similar models that can distinguish between psychiatric disorders.
The brain is a difficult organ to investigate on a cellular level in living human participants, but the brains of deceased patients with psychotic disorders have been investigated in postmortem brain studies.These studies have found significantly increased densities of CD3 1 T lymphocytes and CD20 1 B lymphocytes in areas like the hippocampus and parahippocampus in brains from patients with schizophrenia compared with those from healthy control participants (38,39), which could imply that neuroinflammation may play a role in the pathogenesis of schizophrenia.

Strengths and Limitations
While it is the most comprehensive systematic review and meta-analysis on circulating immune cells in psychotic disorders reported to date, some limitations should be considered.First, the laboratory methods used to identify and categorize the immune cells varied to a high degree across studies, including methods like microscopy, hematology analyzers, and flow cytometry analysis.Flow cytometry has greatly improved our ability to investigate specialized subsets of immune cells, but the method has also undergone considerable development since its invention.Studies that reported results as percentages often did not report the denominator for the percentage, thereby hampering comparisons between studies.Our analyses addressed this by using a standardized effect measure (the SMD) and by focusing on counts rather than percentages of cells.Furthermore, only articles that had undergone the peer-review process were eligible to be included in our study.Additionally, most of the included studies showed bias when it came to the comparability of the patient and the control groups.Less than half of the included studies matched their groups on age and sex, and the studies rarely matched their groups or corrected their results for potential confounding Meta-Analyses of Immune Cells in Psychotic Disorders Biological Psychiatry factors like smoking or obesity (body mass index), both of which are more common in psychotic disorders like schizophrenia (40,41) and are known to influence the immune system (42)(43)(44).Positive and Negative Syndrome Scale scores could be used in subgroup analyses to associate outcomes with symptom severity at the time of measurements; however, the number of studies was insufficient to allow us to perform subgroup analyses on different stages or subtypes of psychotic disorders.Patient setting may be a suboptimal proxy for symptom severity, but it is still useful because it reflects the broad symptom picture of the patients, and it was reported in most studies.Furthermore, the level of heterogeneity between studies was generally high for many of the outcomes, suggesting differences in study populations, control groups, measurement methods, and laboratory techniques.It should also be noted that 2 of the included studies (45,46) had much larger sample sizes than the other included studies, but the leave-one-out sensitivity analyses showed that omitting those studies did not change the results except for the platelet/lymphocyte ratio.More studies reporting results on the Systemic Immune-Inflammation Index (47,48) have been published since the search ended and have not been included in the meta-analysis.

Conclusion and Perspectives
The findings of this meta-analysis indicate that the immune system is activated to a higher degree on average in patients with psychotic disorders than in healthy control participants.Counts of immune cells from both the myeloid (leukocytes, granulocytes, neutrophil granulocytes, and monocytes) and the lymphoid line (B lymphocytes) were found to be significantly elevated compared with healthy control participants, which implies a broad activation of the immune system.However, it should be noted that over half of the included studies did not match or adjust for potential confounding factors such as age, sex, smoking, body mass index, and race, which is recommended in all future studies.Moreover, it is recommended that all future studies collect complete blood count data and calculate immune cell-related ratios routinely on their psychiatric patients because this is an apparently reliable index of inflammation, as well as repeat the test at a specific interval posttreatment that is timed to coincide with treatment response and symptom resolution.Comparing pre-and posttreatment indices of inflammation may also be established as a biomarker of response along with resolution of the inflammatory process.For CSF analyses, there were only 3 studies that could be included, and they showed a nonsignificant trend toward increased white cell count in the CSF; however, more detailed immune cell analyses in the CSF are currently lacking.In-depth CSF analyses are needed to understand whether cellular inflammation in the brain plays a role in psychotic disorders, which could potentially lead to new forms of treatment and diagnostic parameters centered around the immune system.
(or similar headings) or text words: (Psychotic disorders OR schizophrenia OR schizoaffective OR psychotic OR schizophren* OR psychosis) AND (Leukocytes OR Leukocyte count OR Leukocyte* OR white blood cell* OR WBC OR complete blood count OR CBC OR peripheral blood mononuclear cell* OR PBMC OR cell count* OR cell* OR blood-brain barrier OR WCC OR neutrophil* OR basophil* OR "eosinophil* OR monocyte* OR "lymphocyte* OR T lymphocyte* OR T cell* OR T helper OR T regulatory OR T cytotoxic OR Th1 OR Th2 OR Th17 OR B lymphocyte* OR B cell* OR Natural Killer OR CD3 OR CD4 OR CD8 OR CD14 OR CD16 OR CD19 OR CD20 OR CD25 OR CD56) AND (Blood OR Plasma OR Serum OR Spinal puncture OR Lumbar puncture OR Spinal tap OR Cerebrospinal fluid OR CSF).

Figure 1 .
Figure 1.Graphical abstract of the main findings.(A) An overview of the immune cell counts in the blood that have been investigated in $2 studies.(B) Immune cell counts investigated in $2 studies and that are significantly elevated in patients with psychosis compared with healthy control participants.(C) Ratios investigated in $2 studies and that are significantly elevated in patients with psychosis compared with healthy control participants.NK, natural killer; Nr, number; PMBC, peripheral mononuclear blood cell; SMD, standardized mean difference.

Immune cell ra os that were significantly elevated in psychosis compared to healthy controls
0558, p < 0.01 Forest plot of the major leukocyte subset counts in the blood of patients with psychosis compared with healthy control participants.

Lymphocyte ratio Monocyte/Lymphocyte ratio Platelet/Lymphocyte ratio Helper T lymphocyte/Cytotoxic T lymphocyte ratio Random effects model Random effects model Random effects model Random effects model
Forest plot of the B lymphocyte count in the blood of patients with psychosis compared with healthy control participants.
Meta-Analyses of Immune Cells in Psychotic Disorders 336 Biological Psychiatry September 1, 2024; 96:331-341 www.sobp.org/journal Figure 5. Forest plot of the ratios between immune cells in the blood of patients with psychosis compared with healthy control participants.

Table 1 .
Overview of the Results for Immune Cell Counts, Percentages, and Ratios in the Blood of Patients With Psychosis Compared With Healthy Control Participants Investigated in at Least Two Studies