Blood-based immune biomarkers and functional outcomes in acute schizophrenia: A retrospective cohort study

Abstract

Background

Schizophrenia is a chronic psychiatric disorder characterized by acute relapses and significant functional impairment. While antipsychotic medications are effective in managing acute episodes, many patients fail to achieve functional remission. Recent research suggests that immune mechanisms may influence treatment outcomes, yet studies focusing on the relationship between immune biomarkers and functional recovery are limited.

Objective

This study aims to evaluate whether blood-based immune biomarkers can predict functional response and remission in patients with acute schizophrenia.

Methods

A retrospective cohort study was conducted with 354 inpatients diagnosed with schizophrenia, admitted to an acute psychiatric unit between January 2010 and December 2020. Sociodemographic, clinical, and immune biomarker data were extracted from electronic records. Functional outcomes were measured using Global Assessment of Functioning (GAF) scores at admission and discharge. Immune biomarkers assessed included white blood cell counts, ratios and C-reactive protein. Functional response was defined as a GAF score improvement >40 points, while functional remission was defined as a GAF score ≥70 at discharge.

Results

Higher leukocyte counts increased the risk of non-functional response, while a higher platelet–lymphocyte ratio (PLR) was a protective factor. Additionally, Higher lymphocyte and platelet counts were protective against non-functional remission. However, the predictive performance for both functional response and remission was limited, with an AUC ranging from 0.53 to 0.61.

Conclusion

Immune biomarkers, particularly leukocyte counts, PLR, and lymphocyte counts, show significant associations with functional outcomes in acute schizophrenia. However, their predictive value for clinical practice remains limited.

Introduction

Schizophrenia is a chronic psychiatric disorder affecting approximately 1% of the global population and typically manifests during an individual's most productive years.¹ The condition is characterized by recurrent acute relapses and a progressive decline in functional capacity. Although several antipsychotic medications have demonstrated efficacy in managing acute psychotic episodes, up to two-thirds of patients fail to achieve symptomatic remission, and an even larger proportion remain functionally impaired.^2–4The underlying pathophysiology of schizophrenia remains incompletely understood; however, increasing evidence suggests that immune and inflammatory mechanisms may contribute to disease development and progression.⁵ This hypothesis is supported by epidemiological studies,⁶ genetic findings,^7–9 and biomarker research.^10,11 In addition, dysregulation of inflammatory pathways has been proposed as a potential mechanism underlying treatment resistance, highlighting the need to identify immunological biomarkers that may enable early detection of poor treatment response.¹²

Meta-analyses have shown that individuals with schizophrenia exhibit higher levels of white blood cell (WBC) counts, inflammatory cell ratios, and C-reactive protein (CRP) compared with healthy controls.^10,11,13,14 These alterations are detectable from the early stages of the disorder,^15,16 persist during stable phases, and may further increase during acute episodes.^10,17 Elevated inflammatory biomarkers have also been associated with poorer clinical outcomes in schizophrenia.^18–21 Regarding acute treatment response, several baseline biomarkers have been investigated as predictors of response to antipsychotic therapy. A recent narrative review identified 12 studies examining this association, reporting significant but heterogeneous correlations depending on the specific antipsychotic agent studied.²²

Despite these findings, research examining the relationship between immune biomarkers and functional outcomes, rather than solely symptomatic improvement, remains limited. Most available studies focus on clinically stable populations.^23–26 Current treatment guidelines emphasize that improving patient functioning during the acute phase of treatment should be a primary goal in schizophrenia management.^27–29 Nevertheless, many patients continue to experience functional impairment due to persistent negative and cognitive symptoms, residual psychotic symptoms, and adverse effects of antipsychotic treatment.^30–33 In this study, we aimed to evaluate whether blood-based immune biomarkers can predict functional response and functional remission in patients with acute schizophrenia, assess their discriminatory performance for clinical practice, and explore potential differences according to antipsychotic treatment.

Methods

Study design and setting

This register-based retrospective cohort study included individuals with schizophrenia admitted to the acute psychiatric unit of Santa Maria University Hospital in Lleida, Spain. Sociodemographic, clinical, functional, and peripheral immune biomarker data were extracted from electronic medical records. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting cohort studies (Supplementary data).³⁴

Eligibility criteria

Fig. 1 illustrates the patient selection process.

Fig. 1.

Patient selection flowchart for the study.

Inclusion criteria were: (1) admission to the inpatient psychiatric unit between January 1, 2010, and December 31, 2020; (2) age ≥18 years; (3) primary diagnosis of schizophrenia (F20) according to the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM); (4) availability of Global Assessment of Functioning (GAF) scores at both admission and discharge.³⁵

Exclusion criteria were: (1) missing data for total or differential WBC counts, platelets, or CRP; (2) acute or chronic medical conditions that may significantly influence inflammatory parameters, including acute infections, inflammatory or autoimmune diseases, recent surgery or trauma, active malignancy, or other systemic illnesses; (3) pregnancy or breastfeeding; (4) readmission within 30 days of discharge to avoid duplicate observations.

Variables

The following sociodemographic variables were collected: sex (male/female), age (years), type of episode (first-episode or multi-episode schizophrenia), illness duration (years), cardiovascular risk factors (CVRF), including hypertension, dyslipidemia, diabetes mellitus, obesity, and metabolic syndrome defined according to International Diabetes Federation criteria,³⁶ addictive behaviors were also recorded, including use of tobacco, alcohol, cannabinoids, cocaine, opioids, amphetamines, and polysubstance use (defined as use of ≥3 substances excluding tobacco and methadone), as well as problematic gambling. Family history of severe mental disorders was also documented (yes/no).

Treatment at discharge was recorded and reflected the antipsychotic regimen considered effective at the end of hospitalization. Antipsychotics were categorized into seven mutually exclusive groups: risperidone, paliperidone, aripiprazole, olanzapine, clozapine, other second-generation antipsychotics, and first-generation antipsychotics. Prescribed daily doses were converted into chlorpromazine (CPZ) equivalents according to international consensus guidelines.³⁷ For patients receiving antipsychotic polypharmacy, classification was based on the antipsychotic with the highest CPZ-equivalent dose. Additional treatments were also recorded, including antidepressants (yes/no), mood stabilizers (yes/no), and electroconvulsive therapy (yes/no).

Blood samples were collected by trained nursing staff within the first 24 h after admission between 8:00 and 10:00 AM after overnight fasting. The immune biomarkers evaluated included total and differential WBC counts (leukocytes, neutrophils, basophils, eosinophils, monocytes, and lymphocytes), platelet counts, and CRP. The following inflammatory ratios were calculated: neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), platelet-to-lymphocyte ratio (PLR), and basophil-to-lymphocyte ratio (BLR).³⁸ Total and differential WBC counts were measured using flow cytometry with a Sysmex XN analyzer. CRP levels were determined using an immunoturbidimetric assay on a Beckman Coulter automated analyzer.

Outcomes

Functional status was assessed using GAF scores at admission and discharge.³⁹ Two outcomes were defined:

• a)

  Functional response, defined as both a quantitative variable, calculated as the difference in GAF scores between admission and discharge, and a categorical variable defined as a GAF increase of >40 points during hospitalization. Although no standardized cutoffs exist for defining functional response during acute episodes, previous research has demonstrated a strong correlation between the GAF and the Positive and Negative Syndrome Scale (PANSS).⁴⁰ Given that a 30% reduction in PANSS scores is commonly used to define clinical response,⁴¹ we applied equipercentile linking to translate this threshold to the GAF scale. Based on this approach, an increase of 40 points in the GAF score was used as the criterion for functional response in this study.

• b)

  Functional remission, defined as a GAF score ≥70 at discharge, indicating at most mild impairment in social, occupational, or academic functioning.³⁵

Statistical analysis

Statistical analyses were conducted using IBM SPSS Statistics for Windows, version 23 (IBM Corp., Armonk, NY, USA). Continuous variables are expressed as mean ± SD, whereas categorical variables are summarized as frequencies and percentages. Normality of data distribution was assessed using the Kolmogorov–Smirnov test. Because WBC counts, inflammatory ratios, and CRP levels were not normally distributed, these variables were log-transformed before correlation and regression analyses following the approach described by Feng et al. (2013). Comparisons between responders and non-responders and between remitters and non-remitters were performed using Chi-square tests and Student t tests. When assumptions for parametric tests were not met, Fisher's exact tests and Mann–Whitney U tests were applied.

To evaluate functional response, a multivariable generalized linear regression analysis was performed using the change in GAF score from admission to discharge as the dependent variable, with immune biomarkers entered as independent variables. In addition, multivariable logistic regression was used to examine the association between immune biomarkers and functional response, defined as a GAF improvement of more than 40 points. For functional remission, defined as a GAF score > 70 at discharge, a separate multivariable logistic regression analysis was performed to assess the predictive value of immune biomarkers. All models were adjusted for variables known to influence immune parameters, including age, sex, and metabolic syndrome, as well as for significant variables identified in univariate analyses, such as GAF score at admission. These covariates included type of episode and polysubstance use for functional response, and opioid use for functional remission.

Results are reported as standardized β coefficients with 95%CI, for linear regression analyses and as OR with 95%CI, for logistic regression analyses. The Benjamini–Hochberg procedure was applied to account for multiple comparisons, and corrected p values are reported as pcorr. Statistical significance was defined as pcorr < .05.

Receiver operating characteristic (ROC) curve analysis and area under the curve (AUC) values were used to evaluate the discriminatory performance of immune biomarkers for identifying functional response and functional remission.

Results

Demographic and baseline clinical characteristics

A total of 354 inpatients with acute schizophrenia were included in the study (age, 40.6 ± 12.6 years; 30.5% women; illness duration = 11.5 ± 10.1 years). At discharge, 28.2% of patients achieved a functional response, and 24% achieved functional remission. Risperidone was the most commonly prescribed antipsychotic (36.7%), followed by paliperidone (24.2%). Baseline demographic and treatment characteristics are detailed in Table 1.

Table 1.

Sociodemographic, clinical and treatment characteristics of the sample.

	Total sampleN = 354	Functional response (GAF change ≥ 40 points)		P	Functional remission (GAF ≥ 70 at discharge)		P
		Non-respondersN = 254 (71.8%)	RespondersN = 100 (28.2%)		Non-remittersN = 269 (76%)	RemittersN = 85 (24%)
Female, n (%)	108 (30.5)	80 (31.4)	28 (28)	.520	89 (33.1)	19 (22.4)	.061
Age (years), mean (SD)	40.6 (12.6)	41.4 (12.8)	38.7 (12.1)	.067	41.3 (13.3)	38.6 (10)	.096
First episode of schizophrenia, n (%)	36 (10.2)	18 (7)	18 (18)	.002*	26 (9.7)	10 (11.8)	.577
Illness duration, mean (SD)	11.5 (10.1)	11.6 (9.6)	11.4 (11.5)	.838	11.9 (10.8)	10.3 (7.8)	.191
Cardiovascular risk factors, n (%)
High blood pressure	124 (35)	92 (36.2)	32 (32)	.454	94 (34.9)	30 (35.3)	.953
Diabetes mellitus	59 (16.7)	44 (17.3)	15 (15)	.598	47 (17.5)	12 (14.1)	.469
Dyslipidaemia	151 (42.7)	113 (44.4)	38 (38)	.266	112 (41.6)	39 (45.9)	.490
Obesity	112 (31.6)	86 (33.8)	26 (26)	.152	90 (33.5)	22 (25.9)	.191
Metabolic syndrome	132 (37.3)	94 (37)	38 (38)	.862	100 (37.2)	32 (37.6)	.937
Addictive behaviors, n (%)
Tobacco	230 (65)	166 (65.3)	64 (64)	.810	171 (63.6)	59 (69.4)	.325
Alcohol	102 (28.8)	70 (27.5)	32 (32)	.406	79 (29.4)	23 (27.1)	.682
Cannabinoids	193 (54.5)	132 (51.9)	61 (61)	.124	147 (54.6)	46 (54.1)	.932
Cocaine	49 (13.8)	31 (12.2)	18 (18)	.155	37 (13.8)	12 (14.1)	.933
Opioids	21 (5.9)	15 (5.9)	6 (6)	.973	21 (7.8)	0 (0)	.003*
Amphetamines	18 (5.1)	11 (4.3)	7 (7)	.303	14 (5.2)	4 (4.7)	.558
Polysubstance use (3 substance excluding tobacco and methadone)	36 (10.2)	20 (7.8)	16 (16)	.023*	27 (10)	9 (10.6)	.884
Problematic Gambling	3 (0.8)	3 (1.1)	0 (0)	0.368	3 (1.1)	0 (0)	.438
Family history of SMI, n (%)	132 (37.3)	89 (35)	43 (43)	0.163	97 (36.1)	35 (41.2)	.395
Global Assessment of Functioning (GAF), mean (SD)
GAF at admission	29.3 (10.1)	31.9 (9.5)	22.4 (7.9)	<.001*	27.1 (9.1)	36.0 (10)	<.001*
GAF at discharge	59.1 (11.2)	56.7 (11.3)	65.3 (8.4)	<.001*	54.9 (9.2)	72.7 (4)	<.001*
GAF change	29.8 (10.6)	24.8 (7.3)	42.9 (4.8)	<.001	27.8 (9.9)	36.7 (9.9)	<.001*
Antipsychotics, n (%)
Risperidone	130 (36.7)	98 (38.6)	32 (32)	.247	99 (36.8)	31 (36.5)	.956
Paliperidone	86 (24.2)	64 (25.2)	22 (22)	.528	63 (23.4)	23 (27.1)	.495
Aripiprazole	30 (8.4)	18 (7)	12 (12)	.135	24 (8.9)	6 (7.1)	.591
Other second-generation AP	74 (20.9)	48 (18.9)	26 (26)		54 (20.1)	20 (23.5)
First-generation AP	34 (9.6)	26 (10.2)	8 (8)	.520	29 (10.8)	5 (5.9)	.388
Antidepressant, n (%)	43 (12.1)	36 (14.1)	7 (7)	.063	33 (12.2)	10 (11.7)	.902
Mood stabilizer, n (%)	39 (11)	29 (11.4)	10 (10)	.701	34 (12.6)	5 (5.8)	.083
ECT, n (%)	9 (2.5)	8 (3.1)	1 (1)	.225	8 (3)	1 (1.2)	.321

Bold values denote statistical significance after adjusting for multiple comparisons using the Benjamini–Hochberg method.

ECT, electroconvulsive therapy; SMI, serious mental illness.

Immune biomarkers and functional response

Patients who did not achieve a functional response at discharge had higher baseline leukocyte counts (8.08 ± 2.52 vs 7.48 ± 2.10; p = .037) and lower PLR values (113.6 ± 28.9 vs 124.8 ± 34.5; p = .028) (Table 2).

Table 2.

Immunological parameters by functional response and remission status.

Immunological parameters	Total sampleN = 354	Non-respondersN = 254 (71.8%)	RespondersN = 100 (28.2%)	P	Non-functional remissionN = 269 (76%)	Functional remissionN = 85 (24%)	P
Leukocytes (10×9/L)	7.92 (2.44)	8.08 (2.54)	7.48 (2.1)	.037*	7.83 (2.53)	8.17 (2.1)	.264
Neutrophils (10×9/L)	4.76 (2.01)	4.88 (2.14)	4.43 (1.58)	.056	4.75 (2.11)	4.76 (1.65)	.951
Lymphocytes (10×9/L)	2.28 (0.81)	2.31 (0.8)	2.19 (0.84)	.210	2.20 (0.79)	2.51 (0.84)	.002*
Monocytes (10×9/L)	0.66 (0.23)	0.65 (0.22)	0.65 (0.23)	.837	0.65 (0.23)	0.66 (0.19)	.715
Eosinophils (10×9/L)	0.18 (0.13)	0.18 (0.12)	0.17 (0.12)	.252	0.18 (0.13)	0.18 (0.11)	.983
Basophils (10×9/L)	0.04 (0.02)	0.042 (0.026)	0.038 (0.018)	.079	0.042 (0.023)	0.040 (0.016)	.977
Platelets (10×9/L)	234.4 (58.2)	234.4 (59.4)	234.3 (55.5)	.992	230.6 (59.3)	246.4 (53.1)	.027*
NLR	2.36 (1.43)	2.37 (1.47)	2.34 (1.31)	.836	2.43 (1.51)	2.14 (1.1)	.097
MLR	0.313 (0.142)	0.307 (0.135)	0.328 (0.16)	.218	0.322 (0.15)	0.285 (0.11)	.037*
PLR	116.8 (53.9)	113.6 (28.9)	124.8 (34.5)	.028*	118.1 (32.6)	112.5 (37.9)	.408
BLR	0.019 (0.010)	0.019 (0.011)	0.018 (0.009)	.441	0.019 (0.01)	0.017 (0.01)	.153
CRP (mg/L)	7.3 (12.6)	7.7 (12.7)	6.1 (11.9)	.280	7.3 (11.4)	7.4 (15.6)	.941

Bold values denote statistical significance after adjusting for multiple comparisons using the Benjamini–Hochberg method.

NLR, neutrophil-to-lymphocyte ratio; MLR, monocyte-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; BLR, basophil-to-lymphocyte ratio; CRP, C-reactive protein.

In the linear regression model assessing change in GAF score, leukocyte and neutrophil counts were inversely associated with GAF change (β = −0.139; 95%CI, −1.063 to −0.143, p(corr) = 0.0375; β = −0.118; 95%CI, −1.177 to −0.066, p(corr) = 0.0382, respectively), whereas PLR was positively associated with GAF change (β = 0.172; 95%CI, 0.012 to 0.055, p(corr) = 0.015) (Table 3).

Table 3.

Associations between blood-based immune biomarkers and functional outcomes in acute schizophrenia: regression coefficients, odds ratios, and discriminatory performance.

Immunological parameters	GAF change	Non-functional response(GAF change < 40 points)		Non-functional remission(GAF at discharge < 70)
	Beta (95%CI)^a	OR (95%CI) ^b	AUC	OR (95%CI)^c	AUC
Leukocytes (10×9/L)	−0.139 (−1.063 to −0.143)*	1.134 (1.019 to 1.262)*	0.562	0.961 (0.869 to 1.062)	–
Neutrophils (10×9/L)	−0.118 (−1.177 to −0.066)*	1.132 (0.996 to 1.286)	–	1.001 (0.884 to 1.133)	–
Lymphocytes (10×9/L)	−0.096 (−2.669 to 0.198)	1.204 (0.900 to 1.611)	–	0.669 (0.489 to 0.917)*	0.617
Monocytes (10×9/L)	−0.054 (−7.370 to 2.440)	1.113 (0.403 to 3.075)	–	1.010 (0.342 to 2.988)	–
Eosinophils (10×9/L)	−0.075 (−14.931 to 2.665)	3.032 (0.454 to 20.237)	–	1.544 (0.217 to 10.984)	–
Basophils (10×9/L)	−0.105 (−97.997 to 0.068)	16.932 (0.314 to 854.621)	–	0.3152 (0.001 to 18.765)	–
Platelets (10×9/L)	0.038 (−0.013 to 0.027)	1.000 (0.996 to 1.004)	–	0.993 (0.989 to 0.998)*	0.589
NLR	0.011 (−0.709 to 0.875)	1.017 (0.864 to 1.199)	–	1.156 (0.939 to 1.421)	–
MLR	0.081 (−1.866 to 13.805)	0.381 (0.081 to 1.785)	–	8.414 (0.949 to 74.592)	–
PLR	0.172 (0.012 to 0.055)*	0.995 (0.990 to 0.999)*	0.532	1.000 (0.995 to 1.005)	–
BLR	−0.032 (−135.1 to 72.1)	4.113 (0.013 to 25.237)	–	4.113 (0.013 to 25.237)	–
CRP (mg/L)	−0.058 (−0.138 to 0.040)	1.012 (0.990 to 1.034)	–	0.997 (0.977 to 1.017)	–

a

Adjusted for age, sex, metabolic syndrome and GAF at baseline.

b

Adjusted for age, sex, metabolic syndrome, GAF at baseline, type of episode, and polysubstance use.

c

Adjusted for age, sex, metabolic syndrome, GAF at baseline and opioid use.

*

Denote statistical significance after adjusting for multiple comparisons using the Benjamini–Hochberg method.

NLR, neutrophil-to-lymphocyte ratio; MLR, monocyte-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; BLR, basophil-to-lymphocyte ratio; CRP, C-reactive protein.

In logistic regression analyses for non-response (Table 3; Fig. 2), higher leukocyte counts were associated with an increased risk of non-response (OR, 1.134; 95%CI, 1.019–1.262, p(corr) = 0.022), whereas higher PLR was associated with a protective effect (OR, 0.995; 95%CI, 0.990–0.999, p(corr) = 0.016). Neutrophil counts did not show statistically significant associations.

Fig. 2.

Forest-plot of associations between blood-based Immune biomarkers and functional outcomes in acute schizophrenia. Note: *denotes statistical significance.

Regarding the discriminatory performance of immune biomarkers, the optimal leukocyte threshold was approximately 7.32, yielding a sensitivity of 56.7%, specificity of 47.0%, and an AUC of 0.562. For PLR, the threshold was approximately 114.8, with a sensitivity of 42.1%, specificity of 38.6%, and an AUC of 0.532 (Table 3).

Immune biomarkers and functional remission

Patients who did not achieve functional remission had lower baseline lymphocyte counts (2.20 ± 0.79 vs 2.51 ± 0.84, p(corr) = 0.002) and platelet counts (230.6 ± 59.3 vs 246.4 ± 53.1, p(corr) = 0.027), but higher monocyte-to-lymphocyte ratio (MLR) values (0.322 ± 0.15 vs 0.285 ± 0.11) (Table 2).

In the logistic regression model (Table 3; Fig. 2), higher lymphocyte and platelet counts were associated with a lower risk of non-remission (OR, 0.669; 95%CI, 0.489–0.917, p(corr) = 0.012; OR, 0.993; 95%CI, 0.989–0.998, p(corr) = 0.005, respectively).

When assessing discriminatory performance, the optimal threshold for lymphocyte count was 2.37, with a sensitivity of 56.5%, specificity of 36.4%, and an AUC of 0.617. For platelet count, the threshold was 119.5, with a sensitivity of 87.1%, specificity of 68.3%, and an AUC of 0.589 (Table 3).

Impact of antipsychotic treatment on immune biomarkers and functional outcomes

An exploratory analysis was conducted to assess the association between antipsychotic treatments, immune biomarkers, and functional outcomes (Supplementary data).

Functional response

Among patients treated with risperidone, PLR was positively associated with change in GAF score (β = 0.203; 95%CI, 0.040–0.075, p(corr) = 0.028). In patients receiving paliperidone, leukocyte (β = −0.249; 95%CI, −1.616 to −0.105, p(corr) = 0.0382), neutrophil (β = −0.232; 95%CI, −2.018 to –0.065, p(corr) = 0.0427), and basophil counts (β = −0.330; 95%CI, −199.99 to −43.73, p(corr) = 0.015) were inversely associated with GAF change. In the first-generation antipsychotic group, platelet counts were positively associated with GAF change (β = 0.675; 95%CI, 0.023–0.214, p(corr) = 0.045). Among patients treated with aripiprazole, monocyte (β = −0.453; 95%CI, −46.332 to −5.969, p(corr) = 0.013) and eosinophil counts (β = −0.426; 95%CI, −92.096 to −7.549, p(corr) = 0.023) were inversely associated with GAF change. No other significant associations were observed for the remaining antipsychotic groups (Table S1).

In logistic regression analyses, within the risperidone group, higher PLR was associated with a lower risk of non-response (OR, 0.991; 95%CI, 0.983–0.999, p(corr) = 0.022). In the first-generation antipsychotic group, higher lymphocyte counts were associated with an increased risk of non-response (OR, 5.767; 95%CI, 1.121–29.666, p(corr) = 0.011), whereas higher PLR was associated with a reduced risk (OR, 0.986; 95%CI, 0.972–0.999, p(corr) = 0.041). No other significant associations were observed in the remaining antipsychotic groups (Table S2).

Functional remission

For functional remission, within the risperidone group, higher platelet counts were associated with a lower risk of non-remission (OR, 0.991; 95%CI, 0.984–0.999, p(corr) = 0.023). In the olanzapine group, higher neutrophil counts were associated with a lower risk of non-remission (OR, 0.186; 95%CI, 0.051–0.687, p(corr) = 0.012). No other significant associations were identified for the remaining antipsychotic groups (Table S3).

Discussion

This retrospective cohort study aimed to explore the relationship between blood-based immune biomarkers at admission and functional outcomes, measured using the GAF scale, in patients with acute schizophrenia. The findings revealed significant associations between specific immune biomarkers and both functional response and remission. The key findings were as follows: (1) higher leukocyte and neutrophil counts were associated with poorer functional response, whereas a higher PLR was associated with better functional response; (2) higher leukocyte counts increased the risk of non-functional response, while higher PLR showed a protective effect; (3) higher lymphocyte and platelet counts were protective against non-functional remission; (4) predictions for both functional response and remission demonstrated poor to fair discriminatory performance, with AUC values ranging from 0.53 to 0.61; and (5) the relationship between immune biomarkers and functional outcomes appeared to vary depending on the specific antipsychotic treatment used.

Our findings support previous research linking elevated leukocyte counts and inflammatory markers to worse clinical outcomes in schizophrenia.^20,44,45 However, our study extends these findings by focusing specifically on functional outcomes during acute episodes, rather than solely on symptomatic remission. Unlike previous studies that primarily examined clinically stable patients or focused on symptomatic response prediction,²² our work provides additional evidence on the role of immune biomarkers in predicting functional improvement, a clinically relevant yet underexplored aspect of schizophrenia treatment.

The largest study to date investigating the association between immunological biomarkers and acute clinical outcomes in schizophrenia, which included 2598 patients, reported that higher leukocyte counts predicted poorer response to antipsychotic treatment as measured by PANSS total scores.⁴⁶ Although that study used symptomatic response as the primary outcome, our findings are consistent in showing that leukocyte levels also predict functional response. This is not unexpected, as PANSS and GAF scores have demonstrated strong negative correlations.⁴⁰ Consequently, acute treatment response may be captured through changes in both symptom severity and functional status. We also observed that neutrophil counts predicted functional response, although with a slightly smaller effect size compared with total leukocyte counts. To our knowledge, this association has not been previously reported. These results further support the hypothesis that immune system alterations contribute to the neurobiology of schizophrenia and influence treatment response.⁴⁷ In fact, a recent meta-analysis examining anti-inflammatory treatments for psychotic disorders identified small but significant therapeutic effects.⁴⁸ Importantly, these authors suggested that future studies should stratify patients according to immune alterations to better understand treatment benefits. Biomarkers such as those identified in our study could contribute to such stratification strategies.

The protective role of PLR observed in our study is also noteworthy. While elevated PLR is generally associated with worse outcomes in various medical conditions⁴⁹ our findings suggest a more complex role for this marker in schizophrenia. The positive association between PLR and functional improvement may reflect complex interactions between immune processes, dopaminergic signaling, and treatment response during acute episodes.⁵⁰ In this context, baseline PLR levels may reflect biological heterogeneity among patients rather than treatment effects alone, particularly given that prior antipsychotic exposure before admission could not be systematically accounted for. This interpretation aligns with previous evidence showing that higher PLR at the initial diagnosis of schizophrenia was associated with lower mortality.²⁰ Similarly, a longitudinal study in stable schizophrenia patients found that higher PLR was associated with reduced relapse risk.⁵¹ Together, these findings suggest that PLR may have a potentially protective, albeit complex, role in schizophrenia that warrants further investigation.

Functional remission represents a more demanding outcome than clinical response alone.⁵² Although symptomatic remission generally improves functional outcomes, it does not necessarily translate into full functional recovery. In our study, only 24% of patients achieved functional remission at discharge (GAF ≥ 70), which is consistent with previous studies conducted in similar populations.^53,54 Interestingly, the previously reported association with leukocyte counts was not observed for this outcome, whereas higher lymphocyte and platelet counts emerged as protective factors against non-functional remission.

Our findings also expand upon previous research linking inflammatory biomarkers such as IL-2 and IL-12p70 to functional outcomes in schizophrenia, where higher IL-2 levels have been associated with poorer prognosis.^23,24 However, although IL-2 promotes T-lymphocyte proliferation, our results suggest that higher overall lymphocyte counts are protective. This apparent discrepancy may be explained by differences in study design. Previous studies focused on clinically stable outpatients and specific cytokines, whereas our study examined patients during acute schizophrenia episodes and assessed total lymphocyte counts rather than specific immune cell subpopulations. These results highlight the importance of examining immune biomarkers across different disease stages to better understand their role in functional outcomes.

ROC curve analysis was used to assess the discriminatory capacity of immune biomarkers in predicting functional outcomes. Both response and remission predictions showed poor to fair discriminatory performance, with AUC values ranging from 0.53 to 0.61, which is below the 0.8 threshold generally considered clinically useful.⁵⁵ Furthermore, the absolute difference in leukocyte count between responders and non-responders was small (approximately 0.6 × 10⁹/L), further limiting its clinical applicability. These results suggest that although white blood cell counts and inflammatory ratios may reflect pathophysiological processes in acute schizophrenia, their predictive value for clinical decision-making remains limited. Future studies incorporating longitudinal biomarker measurements, detailed symptom profiling, and larger samples will be required to validate these findings.

Finally, our exploratory analyses suggest that different antipsychotics may influence the relationship between immune biomarkers and functional outcomes in distinct ways. Although long-term antipsychotic treatment has been shown to affect overall psychosocial functioning,⁵⁶ the differential effects of individual drugs remain unclear.^57,58 The retrospective design of our real-world cohort limited our ability to identify patients receiving antipsychotic monotherapy, and the small sample sizes for specific antipsychotics restricted individual drug analyses. Nevertheless, we observed differences between treatments, suggesting that the pharmacological profiles of antipsychotics may differentially influence immune pathways and treatment outcomes. However, this hypothesis should be interpreted cautiously, as only a limited number of patients in our sample received monotherapy. Future studies with larger monotherapy cohorts and detailed immune biomarker assessments will be necessary to clarify these relationships.⁵⁹

Strengths and limitations

The main strengths of this study include the relatively large sample size of 354 patients and the comprehensive assessment of immune biomarkers using standardized blood sampling procedures in a real-world clinical setting. In addition, the use of GAF scores allowed us to evaluate functional outcomes, an important yet frequently underreported dimension of schizophrenia treatment.

However, several limitations should be considered. First, the retrospective observational design may introduce selection and information biases and does not allow causal inferences. Second, immune biomarkers were measured at admission, whereas pharmacological treatment was recorded at discharge, introducing a temporal mismatch that may limit causal interpretation of biomarker–outcome associations. Third, the immune biomarkers evaluated—total and differential WBC counts, platelet counts, and CRP—lack specificity and do not allow detailed analysis of underlying immunological pathways. Future research should incorporate more specific markers, such as cytokine profiles or immune cell subtypes. Fourth, biomarkers were assessed at a single time point, preventing evaluation of longitudinal immune changes during hospitalization. Fifth, symptom-based rating scales were unavailable, which prevented analysis of associations between biomarkers and specific symptom domains. Sixth, treatment response was defined solely based on functional improvement using the GAF scale linked to PANSS scores.⁴⁰ Although this approach is supported by previous studies, it has not been formally validated and may affect the robustness of our findings. Finally, stratification by antipsychotic treatment substantially reduced sample size, and patients receiving polypharmacy were categorized according to the antipsychotic with the highest equivalent dose, complicating interpretation of drug-specific results. Therefore, findings from these subgroup analyses should be interpreted with caution.

Conclusions

This study is the first to examine the relationship between blood-based immune biomarkers at admission and functional outcomes in patients with acute schizophrenia. The findings demonstrate associations between differential white blood cell counts, inflammatory ratios, and functional outcomes, although their discriminatory performance remains limited. Future studies should validate these biomarkers in larger cohorts, particularly in patients receiving antipsychotic monotherapy, and investigate longitudinal immune changes to better understand their role in the pathophysiology and treatment response of schizophrenia.

Authors’ contributors

VL-B and MB were responsible for project conceptualization, data analysis, and original drafting. SP-P, EB-R, MA-P, EN-S, LI-P, CA-P, AJ-M, BR-P, and MF were responsible for study recruitment and data collection. MM supervised the project and reviewed the manuscript. All authors revised the manuscript.

Ethical statement

This study was conducted in full compliance with the principles outlined in the Declaration of Helsinki. The study protocol received approval from the Local Ethics Committee of Santa María University Hospital in Lleida, Spain (Approval Reference: CEIC-2341).

Funding

None declared.

Conflict of interest

VL-B reported receiving financial support for continuing medical education from Adamed, Advanz Pharma, Angelini, Casen Recordati, Exeltis, Janssen, Lundbeck, Neurocrine Biosciences, and Rovi outside the submitted work. M.B. has been a consultant for, received grant/research support and honoraria from, and served on the speakers/advisory boards of, or received honoraria for talks and/or consultancy from, Adamed, Angelini, Casen-Recordati, Exeltis, Ferrer, Janssen, Lundbeck, Neuraxpharm, Otsuka, Pfizer, and Sanofi. M.B. acknowledges support from the Spanish Ministry of Health, Instituto de Salud Carlos III (PI20/01066). MM reported receiving financial support for continuing medical education from Adamed, Advanz Pharma, Angelini, Casen Recordati, Exeltis, Janssen, Lundbeck, Neurocrine Biosciences, and Rovi outside the submitted work. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

.

42, 43.