The mystery of angiographically silent macular oedema due to taxanes.

Taxanes are widely used anticancer agents, produced from the plants of the genus Taxus (yews). One of the rare side-effects caused by taxanes is a bilateral cystoid macular oedema (CMO). The particularity of this type of CMO is that it is angiographically silent showing no leakage or pooling on fluorescein angiography (FA). To date, the mechanism of this oedema has not been clearly understood and existing theories do not explain this phenomenon very well. Our aim was to report a case of paclitaxel-induced CMO and put forward a putative explanation for this occurrence. A 64-year-old woman presented with a 7-month history of progressively decreasing bilateral visual acuity with an apparently normal fundus. At entry her best-corrected visual acuity (BCVA) was 0.4 for far and near OD and 0.5 for far and near OS. Optical coherence tomography (OCT) revealed a CMO with a central thickness of 561 μm OD and 488 μm OS; there were no signs of intraocular inflammation. FA showed no capillary leakage and quasi absent late hyperfluorescence OU. Indocyanine green angiography was within normal limits. Classical CMO treatment was ineffective and only discontinuation of paclitaxel resulted in recovery of a normal macular structure after 4 weeks with an increase of BCVA to 0.9 OD and 1.0 OS. In order to understand the properties of taxane drug-induced cystoid macular oedema (TDICMO) we compared the spectral OCT findings of our case to an inflammation-induced CMO of equal thickness and to a case of multifocal choroiditis. The plane of separation of TDICMO was above the external limiting membrane in both cases. In contrast to inflammation-induced CMO where the four external bands were well identified, there was attenuation of these bands in TDICMO but no disruption of the layers as seen in multifocal choroiditis, indicating that the fluid in TDICMO had a high viscosity producing a shadow underneath. TDICMO most probably originates from retinal pigment epithelium dysfunction by their effect on microtubule functions and not from vascular leakage. The content of the CMO seems to be made up of viscous fluid. As the origin of the CMO is not inflammatory, classical CMO treatments have no effect and only discontinuation of the taxane drug allows reversal of the CMO.