Retinal vasculature in cluster headache

Dear Editor,

We would like to commend Ewering et al. for an interesting paper on temporal retinal nerve fibre layer (RNFL) thinning in cluster headache patients, published in the October issue of Cephalalgia (1). The authors performed optical coherence tomography (OCT) on 107 cluster headache patients (mean age: 48.2 ± 9.9 years, 85% males) and found a significant thinning of the RNFL in the temporal and temporal-inferior areas on both eyes, as well as a thickening in the nasal-inferior area, compared to 65 healthy controls. The thinning was more pronounced on the headache side. Cluster patients who used oxygen – notably only a small number (n = 10) – were also found to have bilateral thinning of the temporal RNFL when compared to non-users and healthy controls. Based on the fact that OCT RNFL peaks seem to correlate with the main branches of the retinal veins and arteries (2), the authors explain their findings on the nasal side as an effect of inter-attack vasodilation, and on the temporal side as a result of oxygen-induced vasoconstriction (1).

We recently examined 30 cluster headache patients (mean age: 50.2 ± 12.6 years, 90% males) using digital retinal photos and retinal analysis software (IVAN, Department of Ophtalmology Visual Science, University of Wisconsin, Madison, WI), calculating the central retinal arteriolar and venular equivalents on both eyes (3). We found significantly smaller retinal vein calibre on the headache side compared to the contralateral side in our patients, and also smaller arteriolar calibre, with the latter not being statistically significant. None of our patients used oxygen, and we regarded the reduced vessel calibre as an expression of reduced total blood flow to the ocular area on the headache side in the remission phase of cluster headache.

The methods used in the two studies are not directly comparable, as one measured the thickness of the RNFL and the other measured the retinal vessel calibre. However, Ewering et al. explain their findings with vascular changes in the retina, which calls for our attention. They do not theorize on what could cause the alleged vasodilation in the nasal areas, but they point out that the vessels have more branches in this area, which in our opinion could account for an increased thickness of the RNFL, even without vasodilation. The vasoconstrictive effect of oxygen varies in the different regions of the retina (4), which may explain why the RNFL is thinner only in the temporal areas. However, the poor correlation between macular thickness and amount of oxygen use (r² = 0.007) and the lack of difference in macular thickness between oxygen users and healthy controls raise doubts regarding oxygen use as a causative factor.

Regional RNFL thinning is also documented in migraine patients (5) who do not use oxygen. In our opinion, these studies suggest that retinal vascular changes are inherent phenomena in some primary headaches, possibly caused by recurring attacks accompanied by fluctuations in ocular blood flow. An ongoing altered upstream autonomic tone may perhaps cause chronic changes in the autoregulated retinal vascular bed (3). This study emphasizes the notion that further research is needed on ocular blood flow and retinal nerve fibre changes in cluster headache.