The immune system and its relation to psychosis is a field with emerging interest. Epidemiological studies have shown that individuals diagnosed with schizophrenia have a higher prevalence of autoimmune diseases and infections compared to the general population (1–3). Alterations in various immune factors have been demonstrated by genetic analyses, and elevated levels of pro-inflammatory cytokines and CRP have been reported in schizophrenia vs. healthy controls (4–7). Of particular interest, recent meta-analyses have reported increased neutrophil count in patients with schizophrenia as compared to controls (8–10). Neutrophil-to-lymphocyte ratio (NLR) is another parameter that has received growing attention with respect to the inflammatory hypothesis of schizophrenia. NLR is frequently used as a marker of the balance between two immune pathways: the innate immune system through neutrophil granulocytes, and the adaptive immune system through the lymphocytes. High NLR has been found to correlate with increased levels of cytokines and CRP, and it is used increasingly in the literature as a proxy for systemic inflammation. A correlation between NLR and pro-inflammatory cytokines has been reported for ovarian cancer (11), liver cirrhosis (12), and laryngeal cancer (13). As a predictor, increased NLR is found to correlate with short- and long-term mortality in cardiac diseases and in cancer (14, 15). Also schizophrenia patients seem to have increased NLR compared to healthy controls (9, 10). However, the causality of increased neutrophil count, NLR and other immune markers is not established, and the potential clinical utility of such findings remain to be determined.

In this review, we focus on published studies that shed light on the etiology of altered neutrophil count and NLR in schizophrenia. As the observed associations between schizophrenia and neutrophils or NLR are far from fully understood, with a limited research database, we have chosen a scoping review approach to examine and map the quantity, the variety, and the characteristics of available research on this topic (16). Scoping reviews are particularly suited for issues that need investigation and understanding of larger concepts, and where the knowledge database is heterogeneous in research aim, methodology, and outcome.

Our main question is: Are changes in neutrophil numbers in schizophrenia related to the disease state or the antipsychotic medication? To touch upon this issue, we have defined several sub-questions to be addressed:

To address these questions, we aimed at including studies that compare FEP to controls and chronic schizophrenia patients, drug-naïve to medicated schizophrenia patients, as well as patients in relapse to patients in remission. In addition, we explored available research data on the changes in neutrophil numbers and NLR and their relation to symptoms score [The Positive and Negative Syndrome Scale (PANSS) score or equivalent]. The variables to be compared were neutrophil count and NLR. The purpose was to map existing literature that can contribute to improve our understanding of the causality of the neutrophil alterations in schizophrenia, as well as to identify potential gaps in the knowledge base.

This scoping review is conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analysis extension for Scoping Reviews (PRISMA-ScR) recommendations (16). PRISMA-ScR provides guidelines for a systematic, reproducible, and transparent method of conducting and reporting scoping reviews. The PRISMA-ScR-checklist is provided in Supplementary Table 1.

The literature search yielded 525 hits in total, after duplicates and non-English articles were removed. An initial parallel abstract screening resulted in 97.1% correspondence in inclusion/exclusion. We excluded 466 studies due to wrong study population (e.g., other patient groups, in vivo and in vitro studies), wrong publication type (e.g., reviews and case reports), and wrong outcome measures (Figure 1). A final screening resulted in 49 articles being proceeded to full-text eligibility assessment and further evaluation. Two could not be retrieved and 27 were discarded because of wrong study design (e.g., cross-sectional case-control studies, intervention studies, mixed diagnostic group, or wrong comparison groups). There was disagreement regarding seven articles. These were clozapine studies, and we decided to exclude them due to clozapine's potential to induce neutropenia. Thirteen articles met all inclusion criteria and were used for the synthesis of this review. An overview of the included studies is displayed in Table 1.

In summary, 12 studies were case-control studies, and one was a randomized controlled trial (RCT). Five of the 13 articles contained longitudinal measurements. Twelve of the included articles were published between 2014 and 2020, the remaining article was published in 1998. The number of participants ranged from 22 to 1,144, with a total of 3,977 study participants. Four studies had <100 participants and seven studies had more than 200 participants.

Table 2 gives an overview of the blood sampling procedures. Seven out of 13 studies clearly stated that blood sampling was done at admittance, however, the time frame differed from within 24 h to 3 days or was not specified. Only five studies confirmed the fasting status of the participants, and only four studies reported the time of day of blood sampling.

To follow up on former studies that have shown increased neutrophil count and NLR in patients with schizophrenia compared to healthy controls, we made a broad systematic approach to determine if the neutrophil numbers or NLR in schizophrenia reflects the disease state or if it is a medication effect. Through different formulated sub-questions, we identified some interesting trends.

We first asked how neutrophil count and NLR relate to the diagnosis of schizophrenia. Four studies reported higher neutrophil counts in FEP or early-onset schizophrenia compared to healthy controls, indicating that this is a presenting feature of untreated disease (23, 25, 26, 30). This is in agreement with three recent meta-analyses (8–10).

We next explored the effect of antipsychotics on neutrophil count and NLR. One recent cross-sectional retrospective study reported higher neutrophil numbers in drug-free schizophrenia patients compared to medicated patients (27), while three other cross-sectional studies reported no difference between medicated, antipsychotic-withdrawn and antipsychotic-naïve schizophrenia patients (24, 26, 31). A possible explanation of this controversy is that neutrophil numbers decrease over time during antipsychotic drug exposure, as demonstrated on short-term compared to long-term antipsychotic treated patients (21) and in longitudinal studies (25, 26). This time-dependent decrease in neutrophil numbers postulates two different explanations: Neutrophil numbers may be directly decreased by antipsychotics, or indirectly by processes involved in the decrease of symptoms.

Finally, we asked how neutrophil count and NLR is associated with symptoms and the state of the disease. NLR significantly decreased when patients changed from relapse to remission (20, 31). Moreover, four of five studies found positive correlation between PANSS total score and neutrophil count or NLR. Of note, in the study that did not report correlation with PANSS (30), the patients had lower total symptoms burden than the patients in the other studies (21–23, 26). This indicates that the neutrophil numbers and NLR may be a state marker for disease severity or disease phase in schizophrenic patients, rather than an effect of antipsychotics. Interestingly, no study found correlation to PANSS negative subscale, although one study reported a negative correlation between NLR and BPRS negative symptoms score in medicated patients (27).

The observed association between neutrophils and symptoms burden is interesting. High neutrophil numbers and NLR is not a specific feature of schizophrenia but has been reported in many psychiatric diagnoses, including bipolar disorder, depression, and obsessive compulsive disorder (33–36). Stress is a trigger for mental illness and can cause increased release of neutrophils from the bone marrow by the effect of stress hormones (37, 38). How increased peripheral neutrophils relate to psychiatric disorders is less well-understood, but it may involve IL-1β-mediated neutrophil invasion and modulation of signaling in the brain [reviewed by (47)].

Since leukocyte levels fluctuate throughout the day, the time of blood sampling in essential when comparing white blood cell levels (39–41). The observed differences in leukocytes between schizophrenia patients and healthy controls in the included articles were of small magnitude, comparable to the difference seen throughout a day (42). Also the fasting status is important when comparing blood cell counts, as neutrophil count has been shown to increase after food intake while lymphocyte count decrease (32).

It is therefore a notable weakness in the literature that the description of the blood sampling procedure is often lacking in the articles, which is quite surprising when this is a major source of potential technical and biological bias. Only four studies reported the blood sampling time and only five studies informed that the study participants were fasting before blood draw. Only one study specified that antipsychotic therapy was not initiated at the time of blood sampling. In addition to this, very few studies contained information on the technical equipment or the laboratory analytical procedure of blood sampling, which allow for possible unrepresentative comparisons between and within the different studies. Neutrophils are fragile cells, and the count may be influenced by methodological variables. A recent study concluded that the neutrophil count was significantly reduced at 1 h after blood draw and when stored at room temperature (43). This study found no effect of blood collection technique and anticoagulant. However, another study showed that neutrophils were better preserved in EDTA compared to heparin and citrate (44). Automated hematology analysers and manual counts by microscopy seem to give quite comparable results, but differ in the count of immature cells (45).

The strengths and limitations of this scoping review are mainly defined by the included studies. The studies were similar with respect to inclusion/exclusion criteria for study participants. Inclusion criteria were usually schizophrenia as defined in the DSM-IV/V or ICD-10. Exclusion criteria included a list of factors or diseases that could trigger an inflammatory response, but with variation in which diseases that were included, and some were brief in the description. There were limited relevant articles available, and the studies were conducted in various ways, with differences in recruitment of study participants, duration of observation, and outcome of interest. In addition, there was large variation in the number of study participants, ranging from 22 to 1144. The number of participants and the study design are important factors when evaluating the strength of the results. Longitudinal studies are more powerful and can help to determine cause and effect relationships, while retrospective studies have more bias and need larger numbers to detect significant differences.

Another challenge when comparing different studies, concerns the patient's medication status. In the articles included in this review, different labels and definitions were used to describe the antipsychotic drug history. Several studies had no information about medication status at all. Among the patients labeled as “medicated,” only one study did comparisons based on length of treatment. None of the studies evaluated recent adherence to antipsychotic therapy. Medication compliance in schizophrenia is a constant challenge, with some studies reporting medication non-adherence to range from 20 to 89% (46). This highly variable compliance constitutes a considerable source of error when investigating antipsychotics effect on inflammatory parameters. Furthermore, the definition of being drug-free from antipsychotics varied. Two studies defined it as being off antipsychotics for more than 2 weeks, one study set the limit at 6 weeks, whereas one study did not specify any time frame. Although all antipsychotics have a half-life far below 2 weeks, it is not possible to exclude that hematological side-effects may persist even longer. Different subclasses of antipsychotics could potentially have different effects on neutrophils and lymphocytes. We also cannot rule out other confounding effects.

The scoping review methodology is also prone to its own limitations, such as the challenge to do a quality assessment. With studies of differing nature, comparisons, and weighting of results are difficult.

The trends identified in this article need to be validated. Firstly, the strength of the association between symptomatology and neutrophil count and NLR should be quantified. Furthermore, the time-course of neutrophil decrease after treatment should be further examined. Secondly, except for one study looking at granulocyte number, there are, to our knowledge, no studies comparing neutrophil count or NLR in responders vs. non-responders. It would be interesting to have studies that evaluate if this simple blood marker can be used to guide treatment of schizophrenia.

In this scoping review, we have a clear focus on methodological and technical issues that may influence the measurements of white blood cells. Further research should pursue standardized methods for blood sampling, ensuring identical timing of the procedure. The procedure should be done at the same time of day, and comparison groups should preferably be recruited in the same period/season. Factors that are known to increase leukocyte levels should be registered and accounted for, such as fasting and sleep disturbances. This would ensure more comparable and representative blood differentials and enhance our understanding of the immune system's role in schizophrenia. Further studies should also investigate whether the elevated neutrophil count and NLR is only representative for a sub-group of psychotic patients and if these patients could have beneficial effect of anti-inflammatory treatment.

This scoping review investigated neutrophil count and NLR as cause or consequence of schizophrenia. We used a broad approach and included both longitudinal studies and studies with various comparison groups (medication, disease state, response, and symptoms) to identify potential causes of increased neutrophil numbers. In addition, we evaluated methodological and technical issues that may influence the measurements of blood cells.

We found that neutrophil count and NLR are increased in FEP and schizophrenia patients independent of antipsychotic use, compared to the healthy controls. A decrease in neutrophil count and NLR is reported after initiation of antipsychotic therapy, however, this may reflect an indirect effect of the parallel decrease of positive symptoms. There are substantial deficiencies in the available research, particularly concerning methodological standards and comparability of blood sampling procedures. Factors that affect the leukocyte count should be accounted for.

Further research should quantify the correlation between neutrophil count or NLR and symptomatology and investigate its clinical utility for monitoring the course of the disease.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

AS, VS, and AT conceptualized the project. AS and AT performed the literature screening and the review process. AS wrote the first draft of the manuscript. VS and AT critically reviewed the manuscript. All authors contributed to writing and revising the manuscript.

This work was supported by the Research Council of Norway (NORMENT CoE funding program; Grant No. 223273) and Helse Vest RHF, Norway (Grant No. F-12544).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.