Inflammatory molecules in tear fluid in retinitis pigmentosa: an exploratory case-control study

This observational case-control study aimed to compare tear-fluid inflammatory mediator profiles between patients with retinitis pigmentosa (RP) and healthy controls (HC), and to explore associations with RP severity, anxiety, daytime sleepiness, and sleep quality. A total of 78 adults with confirmed RP and 32 HC from a reference controlled-environment cohort were enrolled. RP participants completed anxiety, sleep quality, and sleepiness questionnaires. Tear samples were collected under standardized controlled-environment conditions, and biomarker analyses were conducted using one tear sample per participant. A customized multiplex panel was quantified using xMAP/Luminex technology, and substance P (SP) was measured by competitive ELISA. Between-group differences and within-RP associations were initially assessed using multivariable models adjusted for age and sex, with multiplicity controlled using the Benjamini-Hochberg false discovery rate (BH-FDR) in exploratory analyses. Given the age/sex imbalance between groups, additional sensitivity analyses were performed, including restriction to the overlapping age range, sex-stratified analyses, and inverse probability weighting based on age and sex and 1:1 age- and sex- ased matching. RP tear fluid showed a consistent pattern of lower mediator concentrations and reduced detectability compared with controls. In the main quantitative analyses, RP participants had significantly lower levels of EGF, GRO, IL-8, MCP-1, and IL-1RA, and these between-group patterns were directionally consistent across sensitivity analytes. For low-detection analytes, RP participants also showed reduced detectability relative to controls, although these results should be interpreted more cautiously. Within the RP cohort, tear biomarker profiles were broadly similar across clinical severity and anxiety, sleep quality, and sleepiness strata, with no associations remaining significant after BH-FDR correction. These findings suggest that tear-fluid immune mediators may provide complementary, non-invasive information on ocular surface neuroimmune alterations in RP; however, because the control cohort was not prospectively age- and sex-matched and tear-dilution normalization variables were not available, the between-group findings should be interpreted as exploratory and hypothesis-generating.