Complete blood count-derived inflammatory markers and symptom phenotypes in schizophrenia spectrum disorders: A hierarchical clustering analysis

Background: Schizophrenia spectrum disorders (SSDs) show substantial clinical heterogeneity, potentially reflecting distinct pathophysiological mechanisms. Peripheral immune-inflammatory alterations have been implicated in SSDs, yet their links with specific clinical phenotypes remain largely unexplored. This study thus aimed to examine associations between complete blood count-derived inflammatory markers and symptom patterns in inpatients with SSDs. Methods: This cross-sectional study included inpatients with SSDs. Hierarchical clustering was applied to Positive and Negative Syndrome Scale (PANSS) five-factor scores to identify symptom-based phenotypes. Six markers - neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), and aggregate index of systemic inflammation (AISI) - were compared across clusters using analysis of variance and multivariable linear regressions. Results: Among 454 inpatients with SSDs aged 18-65, three clusters emerged: Positive/excited-dominant (n = 185, 40.8%), Negative/disorganised/depressive-dominant (n = 133, 29.3%), and Balanced (n = 136, 30.0%). All inflammatory markers significantly differed across clusters: NLR (F = 3.29, p = 0.038), MLR (F = 11.13, p < 0.001), PLR (F = 3.71, p = 0.025), SII (F = 3.34, p = 0.036), SIRI (F = 7.42, p < 0.001), and AISI (F = 6.49, p = 0.002). The Negative/disorganised/depressive-dominant cluster exhibited the most pronounced inflammatory profile, with covariate-adjusted increases compared to the Balanced cluster of +30% for MLR (p = 0.001), +20% for PLR (p = 0.040), +32% for SII (p = 0.043), +43% for SIRI (p = 0.011), and +52% for AISI (p = 0.015). The Positive/excited-dominant cluster showed selective, covariate-adjusted elevations in MLR (+19%, p = 0.020) and SIRI (+29%, p = 0.028) versus the Balanced cluster. Conclusions: Distinct SSD phenotypes are associated with heterogeneous inflammatory profiles, with monocyte-driven inflammation particularly characterizing negative/disorganised/depressive presentations, probably supporting phenotype-stratified approaches in precision psychiatry.