Abstract
BACKGROUND: Anxiety symptoms in elderly patients with cognitive impairment (CI) often reflect shared neurobiological processes rather than distinct psychiatric disorders. Current diagnostic approaches lack objective biomarkers for early identification. This study investigated the association between serum biomarkers and anxiety disorder status in elderly men with CI and to evaluate the exploratory discriminative ability of cognitive domains and biomarker profiles in differentiating CI patients with and without comorbid anxiety. METHODS: This cross-sectional retrospective study analyzed 86 elderly male CI patients (Group A: CI alone, n = 41; Group B: CI with anxiety, n = 45) at Jiangsu Rongjun Hospital (June-December, 2024). Anxiety disorder diagnosis was established through structured clinical interviews based on the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria conducted by two independent psychiatrists, with the Hamilton Anxiety Scale (HAMA) serving as an initial severity screening instrument. The Montreal Cognitive Assessment (MOCA) was used to evaluate cognitive function. Enzyme-linked immunosorbent assay (ELISA) was used to measure serum Tau protein (Tau), β-amyloid (Aβ), visinin-like protein 1 (VILIP-1), malondialdehyde (MDA), tumor necrosis factor-α (TNF-α), and interleukin-6 (IL-6); reverse transcription-polymerase chain reaction (RT-PCR) quantified microRNA-34c (MiR-34c). Patients with acute inflammation (C-reactive protein [CRP] >10 mg/L) were excluded. Bonferroni correction was used to address multiple comparisons across 25 simultaneous tests, and multivariate regression analysis was controlled for demographic and clinical confounders. Receiver operating characteristic (ROC) analysis was used to determine the discriminative ability. RESULTS: Group B showed worse cognitive performance across the MOCA domains, with attention (area under the curve [AUC] = 0.738) and delayed recall (AUC = 0.742) demonstrating the strongest discriminative ability. Biomarker analysis revealed elevated Tau (AUC = 0.957), MDA (AUC = 0.941), and VILIP-1 (AUC = 0.914) in anxiety patients. Within-group analyses showed that anxiety severity correlated negatively with MiR-34c and positively with Tau, Aβ, MDA, IL-6, and VILIP-1. Under the Bonferroni-adjusted threshold (p < 0.002), only MDA in Group B (r = 0.478, p = 0.001) and MiR-34c in Group B (r = -0.523, p < 0.001) remained significant. Multivariate analysis identified these factors as independently associated with the outcome after controlling for demographics and comorbidities. However, given the substantial baseline imbalances between the groups, these associations should be interpreted with caution. CONCLUSION: Combined cognitive assessment (attention, delayed recall) and serum biomarkers (Tau, MDA, VILIP-1, MiR-34c) demonstrate promising discriminative ability for identifying anxiety in elderly male patients with CI. These findings are exploratory and derived from a single-center cohort of retired male military veterans with pronounced baseline group imbalances, which substantially limits generalizability to the broader elderly CI population. The identified markers may reflect shared neuroinflammatory and oxidative stress pathways underlying both cognitive and emotional dysfunction, warranting further investigation as potential targets for integrated therapeutic approaches. Validation in prospective, multicenter, sex-inclusive cohorts with balanced comparison groups is essential before any clinical application can be considered.