Paclitaxel‐induced cystoid macular oedema

Abstract

A74-year-old Taiwanese woman with stage IVA poorly differentiated ovarian serous carcinoma was referred for a one-month history of decreased vision in both eyes. She had been receiving chemotherapy including paclitaxel 80 mg/m2 and carboplatin AUC 2.0 every week, over the previous three months. She denied any history of previous ocular disease nor intraocular surgery. At presentation, best-corrected visual acuity was 20/50 OU. The anterior segment evaluation revealed no inflammation and bilateral 1 + nuclear scleroses. Dilated funduscopic examination showed diminished foveal reflex and mottled maculae in both eyes, with no epiretinal membranes or vitreomacular traction (Fig. 1A,B). Fluorescein angiography (FA) results were normal until late frames, where faint stellate dye pooling could be identified centrally (Fig. 1C,D). Spectral-domain optical coherence tomography (SD-OCT) (Cirrus HD-OCT Model 4000; Zeiss, Jena, Germany) depicted foveal thickening and intraretinal hyporeflective cystic cavities, located more prominently in the central outer plexiform layer and less in the adjacent inner plexiform layer. The central retinal thicknesses (CRT) were 478 μm OD and 475 μm OS (Fig. 1E,F). As the patient did not take any other drugs associated with cystoid macular oedema (CME), she was diagnosed with paclitaxel-induced CME. 2% dorzolamide three times a day was initially prescribed, trying to relieve her visual symptoms. Nevertheless, CME persisted after 1 month of dorzolamide. After discussion with her gynaecologist, paclitaxel was replaced with gemcitabine. Following discontinuation of paclitaxel for 1 month, the morphology of maculae returned to normal, CRT decreased to 247 μm OD and 249 μm OS, and no cystic spaces were detected (Fig. 1G,H). Her visual acuity returned to 20/40 OD and 20/22 OS. The taxane class of drugs, such as docetaxel, paclitaxel or nanoparticle albumin-bound paclitaxel, are microtubule stabilizing agents used to treat various malignancies. Bilateral CME without fluorescein leakage is a rare complication of these medications. Including our patient, 24 cases of taxane-induced CME were identified in English literature (Teitelbaum & Tresley 2003; Telander & Sarraf 2007; Murphy et al. 2010; Georgakopoulos et al. 2012; Koo & Kim 2012). The onset of CME could start from two months to three and a half years after taxane treatment. Fluorescein angiography (FA) and spectral-domain OCT are crucial for diagnosis and follow-up. Generally, taxane-induced CME had a normal fluorescein angiogram with no or minimal capillary leakage. Similar non-leaking CME can be seen in niacin-related maculopathy, some types of retinitis pigmentosa, juvenile X-linked retinoschisis and Goldmann-Favre syndrome. The pathogenesis of CME without capillary leakage remains unclear. Poor response to anti-inflammatory treatments might exclude the possibility of inflammatory cause. Lack of evidence of fluorescein leakage, and poor response to systemic or intravitreal injection (IVI) of bevacizumab might denote intact blood–retinal barriers. Several authors proposed that direct toxicity of taxane on Müller cells may deteriorate ion regulation and cause cellular swelling and cystic formation (Teitelbaum & Tresley 2003; Georgakopoulos et al. 2012; Koo & Kim 2012). Dysfunctional retinal pigment epithelium (RPE) due to taxane toxicity might be another possibility owing to the malabsorption across the RPE instead of vascular leakage. Other authors proposed that the reasons for non-leaking CME could be that the fluid leaking rate might be too slow to be detected on FA, or that only molecules smaller than fluorescein could selectively leak from capillaries (Telander & Sarraf 2007). Adjuvant ophthalmic treatment remains controversial. Without cessation of taxane, IVI of bevacizumab or triamcinolone, subtenon injection of triamcinolone, topical prednisolone and non-steroidal anti-inflammatory drugs, and topical dorzolamide had little or no efficacy. A few cases had positive treatment effect with oral carbonic anhydrase inhibitors. The only definite treatment for taxane-induced CME is cessation of the offending agent. The improvement could be detected as soon as 3 weeks after discontinuation of taxane drugs. Although taxane-induced CME is reversible, the impact on visual function and clinical antineoplastic effects relying on taxanes should be weighed carefully. The incidence of taxane-related maculopathy may be underestimated, and this ocular complication requires the clinical practitioners to be cautious when using taxanes. Patients on this medication should monitor their vision regularly, and should be referred for ocular examination if vision symptoms develop.