Circular statistics for analyzing changes in retinal nerve fiber layer

Jose M Abalo-Lojo; Manuel Arias-Gomez; Francisco Gonzalez

doi:10.1016/j.jcjo.2024.03.002

Circular statistics for analyzing changes in retinal nerve fiber layer

Can J Ophthalmol. 2024 Apr 3:S0008-4182(24)00065-6. doi: 10.1016/j.jcjo.2024.03.002. Online ahead of print.

Authors

Jose M Abalo-Lojo¹, Manuel Arias-Gomez², Francisco Gonzalez³

Affiliations

¹ Department of Ophthalmology and Health Research Institute of Santiago de Compostela (IDIS), Complexo Hospitalario Universitario de Santiago de Compostela, Santiago de Compostela, A Coruña, Spain; Department of Surgery, University of Santiago de Compostela, Santiago de Compostela, A Coruña, Spain; Center for Research in Molecular Medicine and Chronic Diseases (CIMUS), University of Santiago de Compostela, Santiago de Compostela, A Coruña, Spain.
² Department of Neurology, Complexo Hospitalario Universitario de Santiago de Compostela, Santiago de Compostela, A Coruña, Spain.
³ Department of Ophthalmology and Health Research Institute of Santiago de Compostela (IDIS), Complexo Hospitalario Universitario de Santiago de Compostela, Santiago de Compostela, A Coruña, Spain; Department of Surgery, University of Santiago de Compostela, Santiago de Compostela, A Coruña, Spain; Center for Research in Molecular Medicine and Chronic Diseases (CIMUS), University of Santiago de Compostela, Santiago de Compostela, A Coruña, Spain. Electronic address: francisco.gonzalez@usc.es.

PMID: 38580212
DOI: 10.1016/j.jcjo.2024.03.002

Abstract

Objective: To evaluate the use of circular statistics to analyze retinal nerve fibre layer (RNFL) thickness in eyes with and without a prior history of optic neuritis (ON).

Design: Single-centre consecutive study.

Participants: Twenty-two multiple sclerosis patients and 20 healthy control subjects.

Methods: Data on 28 eyes with a history of ON of 22 multiple sclerosis patients and 40 eyes of 20 healthy control subjects collected in 2010 and 2015. RNFL thickness was measured separately in 12 sectors around the optic nerve head. We used circular statistics to calculate the mean weighted vector of RNFL thickness for each sector and eye in 2 measurements made 5 years apart (2010 and 2015). Comparisons of weighted mean vectors between groups were made using a paired Mardia-Watson-Wheeler test.

Results: The directions of the mean weighted vectors for ON eyes were 45.8º in 2010 and 56.0º in 2015, whereas in control eyes the directions were 319.4º in 2010 and 188.9º in 2015. No significant differences were found between 2010 and 2015 in any of the 2 groups. However, significant differences were found between ON and control eyes in 2010 and 2015.

Conclusions: This paper provides an example of how to use circular statistics in cases of directional data in ophthalmology and demonstrates that circular statistics are a suitable tool for this purpose.