Mid Sweden University

Transmission of upper respiratory tract infections (URTI) has been demonstrated at the ocular surface (Bischoff et al., 2011). Thus, the immunological profile of the tear fluid likely plays an important role in host defence against URTI, and moreover provides a non-invasive medium for assessment of immune status. We recently demonstrated that tear secretory IgA (SIgA) has potential as a biomarker of URTI risk (Hanstock et al., 2016). It is likely that several other antimicrobial proteins abundant in tears such as lactoferrin (Lf) and lysozyme (Lys) contribute to host defence at the ocular surface (McDermott, 2013).

 PURPOSE: To explore the potential of tear Lf and Lys to evaluate risk of subsequent URTI independently and in combination with tear SIgA data from the same subjects presented in Hanstock et al., (2016).

 METHODS: Forty healthy, physically active subjects were recruited during the common-cold season. Subjects reported upper respiratory symptoms (URS) daily and provided weekly tear samples for 4 weeks. If URS were reported for ≥ 48h, subjects provided a nasopharyngeal swab for identification of common-cold pathogens using RT-PCR and a tear sample. Following an episode of URS, subjects reported daily URS until they had been symptom-free for 4 weeks at which time a ‘Recovery’ tear sample was collected. Tear Lf and Lys concentration was determined using ELISA.

 RESULTS: Eleven students reported episodes of URS; nine returned positive virology tests for human rhinovirus (URTI). Twenty-two students remained symptom-free during the monitoring period (Healthy) and seven were excluded due to non-compliance. Tear Lys concentration (Lys-C) and secretion rate (Lys-SR) were lower in URTI vs. Healthy (p<0.01 and p<0.05 respectively) but there was no difference in Lf-C and Lf-SR. The potential of Lf and Lys to assess URS risk was determined by comparing Lf and Lys the week before URS with Recovery samples. Tear Lf-C, Lf-SR and Lys-C were not altered before URS, whereas Lys-SR tended to be reduced in the week before URS (p<0.1). A binary logistic regression incorporating tear flow rate, Lys-C, Lf-C and SIgA-C as predictors was able to correctly identify subjects at risk of URS in the next week with 70% accuracy (95% CI: 54 – 85%) but only 27% sensitivity.

 CONCLUSION: Although tear Lys was decreased during URTI, the new model including tear Lys and Lf was not able to improve upon the utility of tear SIgA alone to assess URS risk in this small cohort. Larger datasets will be required to evaluate and optimise model performance for models based on tear SIgA, possibly in combination with other biomarkers, to predict URTI.