Retinal morphology across the menstrual cycle: insights from the UK Biobank

Figures

• 

  Fig. 1. Flowchart detailing inclusion and exclusion of participants in UKBB. The starting point consisted of women included in either of the two ophthalmic assessments who underwent retinal imaging (OCT and CFP). When participants attended both ophthalmic assessments, the earliest one was preferred. The same participant might be excluded in different categories within each exclusion box. * “Healthy eumenorrheic women”, refers to the final subset of women, excluding participants with specified health conditions and criteria that could interfere with the analysis.

• 

  Fig. 2. Mean oestradiol levels per day of the cycle in our UKBB sample. Oestradiol levels in pmol/L are represented on the curved line of the left y-axis. The number of participants who had hormone readings for each cycle day is shown on the right y-axis. Due to the small sample size, information is limited for individuals on cycle days beyond 29.

• 

  Fig. 3. Distribution of arteriolar and venular perfusion densities and CRAE and CRVE calibres across different days of the menstrual cycle. Participants were grouped based on the reported cycle day, and ocular measurements were initially averaged for each participant’s eyes per day. The top graphs show density (microvasculature), and the bottom graphs show central vessel calibre (macrovasculature). The left-side graphs portray arteriolar parameters, while the right-side graphs present venular measurements. The scale was maintained within both density and calibre graphs to allow comparisons of arterial and venous parameters. Follicular phase measurements are dotted in orange, and luteal phase measurements are dark blue. CRAE Central Retinal Artery Equivalent, CRVE Central Retinal Vein Equivalent, avg average

• 

  Fig. 4. Framework for estimating menstrual cycle phases based on self-reported questionnaire information from the UKBB. Participants were categorised into three groups based on reported cycle length: 21–24 days, 25–30 days, and 31–35 days. Follicular and luteal phases were estimated based on the day of the cycle reported by the participants and the duration of the phases reported by Bull et al.36 in a large epidemiological study based on app-reported cycle data1. The length of each phase reported by Bull et al.36 was used to set thresholds according to the reported cycle length, which were applied to “days since the last cycle.”

• 

  Fig. 5. Retinal imaging and segmentation in UKBB. a Colour Fundus Photograph (CFP) of the posterior pole, centred on the macula and including the optic disc and temporal vascular arcades. Represented in white at the centre are the inner circles of the Early Treatment Diabetic Retinopathy Study (ETDRS) zones, overlaid on a CFP for demonstration purposes. Typically, the ETDRS grid is superimposed on a macula-centred infrared image. C central subfoveal field, IS inner superior field, IT inner temporal field, II inner inferior fields, IN inner nasal field. b Black and white segmentation mask of retinal vessels generated by AutoMorph. c Horizontal B-scan OCT over the fovea. Number 1 points to the top segmentation line in yellow, representing the inner limiting membrane; 2 points to the middle segmentation line on the interface between the outer plexiform and inner nuclear layer, and 3 points to the outer edge of the Retinal Pigment Epithelium (RPE). Inner retinal thickness is measured from line 1 to line 2, outer retinal thickness from line 2 to line 3, and total retinal thickness from line 1 to line 3. OCT segmentation measures were automatically calculated by TABS software.