Retinal Nerve Fibre Layer Thickness And Covariates In The EPIC-Norfolk Cohort

Abstract

 

Purpose:

 

To describe retinal nerve fibre layer (RNFL) measures and covariates in the EPIC-Norfolk cohort (European Prospective Investigation of Cancer).

 

Methods:

 

The third health check of EPIC-Norfolk commenced in 2006 and included a comprehensive ophthalmic assessment. Average RNFL thickness and thickness modulation were measured using scanning laser polarimetry (GDx VCC). Images with a quality score less than 8 were excluded. Other covariables considered were LogMAR visual acuity, axial length (IOLMaster), intraocular pressure (Ocular Response Analyzer), age, sex, body mass index, social class, education, self-reported smoking status and alcohol intake, and physical activity index (questionnaire derived). All covariables were fitted together in two multivariable linear regression models with RNFL thickness and modulation as dependent variables. Analyses were repeated including typical scan score (TSS) in the models. Data were considered from both eyes of each participant, and clustered analyses were carried out to account for the correlation between eyes (generating robust standard errors).

 

Results:

 

Data were analysed from 4,196 eyes of 2,402 participants with a mean age of 67 years. 43% were male. Mean (SD) RNFL thickness was 56.3 (6.3) μm and mean (SD) RNFL modulation was 22.0 (4.9) μm. Older age was independently associated with thinner RNFL and smaller modulation (both p<0.001). Greater axial length was associated with thicker RNFL and a lower RNFL modulation (both p<0.001) in models without TSS as a covariable. However, when TSS was included, the association between axial length and RNFL thickness was no longer significant. Furthermore, adding TSS to the model increased the slope of the negative association between age and RNFL thickness. The table summarises the regression models for RNFL thickness as the dependent variable, with and without TSS as a covariable.

 

Conclusions:

 

Older age was associated with thinner RNFL, which is in keeping with current evidence. Older age was also associated with less RNFL thickness modulation. The apparent association between axial length and RNFL thickness was due to spuriously high RNFL thickness measurements in eyes with low TSS, and that longer eyes were more likely to have low TSS.  

 

Keywords: imaging/image analysis: clinical • nerve fiber layer • aging