BackgroundCerebrospinal fluid (CSF) immune alterations have been associated with mental disorders, neurological disease, and CNS infections; however, comprehensive large-scale longitudinal CSF studies are lacking. MethodsBy using the Clinical Laboratory Information System (LABKA) Research Database in the Central Denmark Region (1994–2012), we included 15,030 individuals tested for CSF WBC, CSF/serum albumin ratio, IgG index, total protein, albumin, or IgG with follow-up for the risk of mental disorders, psychotropic prescriptions, neurological diseases, or CNS infections, estimated by Cox regression. ResultsAmong individuals receiving a mental disorder diagnosis (N = 1,147) after a CSF test, 30·0% had an abnormal CSF test result, while for those with a neurological disease (N = 3,201), 39·9% had abnormal test results, and among individuals with CNS infections (N = 1,276), 73·0% had abnormal test results. Individuals with abnormal CSF test results had an increased risk of mental disorders (HR = 3·20; 95%CI = 2·86-3·59), neurological diseases (HR = 12·40; 95%CI = 11·65-13·20), and CNS infections (HR = 338·59; 95%CI = 299·06-383·35) compared to individuals not registered with a CSF test. However, the risk of mental disorders was higher (P < 0·001) after CSF test results within the normal range (HR = 4·45; 95%CI = 4·08-4·86), whereas for neurological diseases (HR = 9·72; 95%CI = 9·19-10·29) and CNS infections (HR = 55·17; 95%CI = 47·12-64·60), the risk was highest after abnormal CSF test results (all P < 0·001). The risk of organic mental disorders tended to be highest in individuals with abnormal CSF test results (HR = 19·30; 95%CI = 13·44-27·71) even though not significantly different from the risk in the group of individuals with CSF test results in the normal range (HR = 13·55; 95%CI = 9·36-19·60) (P ≥ 0·05). Abnormal CSF test results were associated with an elevated risk of psychotropic prescriptions (HR = 3·91; 95%CI = 3·66-4·18), as were CSF test results within the normal range (HR = 4·26; 95%CI = 4·03–4·51) (P < 0·05). ConclusionsImmunological CSF abnormalities are associated with an increased risk of mental disorders, neurological disease, and particularly CNS infections; however, the included CSF parameters were not specific for mental disorders and the relevant CSF biomarkers in psychiatry are yet to be discovered.

A causal relationship between immune dysregulation and schizophrenia has been supported by genome-wide association studies and epidemiological evidence. It remains unclear to what extent the brain immune environment is implicated in this hypothesis. We investigated the immunophenotype of microglia and the presence of perivascular macrophages and T lymphocytes in post-mortem brain tissue. Dorsal prefrontal cortex of 40 controls (22F:18M) and 37 (10F:27M) schizophrenia cases, of whom 16 had active psychotic symptoms at the time of death, was immunostained for seven markers of microglia (CD16, CD32a, CD64, CD68, HLA-DR, Iba1 and P2RY12), two markers for perivascular macrophages (CD163 and CD206) and T-lymphocytes (CD3). Automated quantification was blinded to the case designation and performed separately on the grey and white matter. 3D reconstruction of Iba1-positive microglia was performed in selected cases. An increased cortical expression of microglial Fcγ receptors (CD64 F = 7.92, p = 0.007; CD64/HLA-DR ratio F = 5.02, p = 0.029) highlights the importance of communication between the central and peripheral immune systems in schizophrenia. Patients in whom psychotic symptoms were present at death demonstrated an age-dependent increase of Iba1 and increased CD64/HLA-DR ratios relative to patients without psychotic symptoms. Microglia in schizophrenia demonstrated a primed/reactive morphology. A potential role for T-lymphocytes was observed, but we did not confirm the presence of recruited macrophages in the brains of schizophrenia patients. Taking in account the limitations of a post-mortem study, our findings support the hypothesis of an alteration of the brain immune environment in schizophrenia, with symptomatic state- and age-dependent effects.