Reliability and validity of conjunctival ultraviolet autofluorescence measurement

Background: Conjunctival ultraviolet autofluorescence (UVAF) photography was developed to detect and characterise pre-clinical sunlight-induced UV damage. The reliability of this measurement and its relationship to outdoor activity are currently unknown. Methods: 599 people aged 16-85 years in the cross-sectional Norfolk Island Eye Study were included in the validation study. 196 UVAF individual photographs (49 people) and 60 UVAF photographs (15 people) of Norfolk Island Eye Study participants were used for intra- and inter-observer reliability assessment, respectively. Conjunctival UVAF was measured using UV photography. UVAF area was calculated using computerised methods by one grader on two occasions (intra-observer analysis) or two graders (inter-observer analysis). Outdoor activity category, during summer and winter separately, was determined with a UV questionnaire. Total UVAF equalled the area measured in four conjunctival areas (nasal/temporal conjunctiva of right and left eyes). Results: Intra-observer (ρ-c=0.988, 95% CI 0.967 to 0.996, p<0.001), and inter-observer concordance correlation coefficients (ρ-c=0.924, 95% CI 0.870 to 0.956, p<0.001) of total UVAF exceeded 0.900. When grouped according to 10 mm 2 total UVAF increments, intra- and inter-observer reliability was very good (κ=0.81) and good (κ=0.71), respectively. Increasing time outdoors was strongly with increasing total UVAF in summer and winter (p trend <0.001). Conclusion: Intra- and inter-observer reliability of conjunctival UVAF is high. In this population, UVAF correlates strongly with the authors' survey-based assessment of time spent outdoors.