SSRI-associated optic neuropathy

Case reports

The following cases were diagnosed as optic neuropathy based on clinical history, physical signs, and following extensive investigation including MRI scan, CXR, FBC, U and E, serum B12, coagulopathies, autoimmune serology, ESR, CRP, serum ACE, fundus fluorescein angiography (FFA), optical coherence tomography, visual fields (VF), electroretinogram (ERG), and visual evoked potentials (VEPs). Leber's mutation and serology for lyme, syphilis and bartonella were also performed where appropriate. Each case has been assessed using the Naranjo algorithm to indicate likelihood of adverse drug reaction.¹ Many additional components from the following case histories have also been summarised in Table 1.

Discussion

In 2004, approximately 7% of UK adults were recieving SSRI treatment for a range of 4.8–7.7 years.² Currently there are no reports of SSRI associated ON, although papilloedema has been reported.³ A potential mechanism for ION has been previously described in relation to raised serotonin levels.^{4, 5} A single case of central retinal vein occlusion exists along with reports of deep vein thrombosis and ischaemic stroke.^{6, 7, 8}

Four out of five cases summarised in Table 1 presented with visual loss from suspected posterior ION (PION). Two cases experienced progressive bilateral VF loss, one altitudinal, the other severe constriction. Two cases presented with sudden severe unilateral LOV. All patients were relatively young with a long history of SSRI treatment. One patient had a typical appearance of bilateral simultaneous anterior ION (AION).

All the cases were able to stop SSRIs without any further deterioration (Table 1). One patient with bilateral progressive LOV recovered all of her vision over 18 months. Two patients with unilateral SLOV did not recover function in the affected eye but retained normal vision in the second eye.

A ‘non-arteritic' ischaemic origin was suspected in all cases. Toxicity was felt to be unlikely as three cases were of sudden onset. Toxic ON is typically characterised by central or cecocentral scotomas. In this series, two cases were unilateral and of those with bilateral involvement, two had altitudinal VF loss and one experienced constricted VF. AION (involving short posterior ciliary arteries) presents with optic disc swelling, often segmental and sudden, associated with altitudinal VF loss.⁹ The clinical features of PION (involving pial branches of central retinal artery) are typically normal disc appearance at the outset with acute, painless LOV in one or both eyes; however, in some cases it can be progressive.⁹ PION in young healthy patients is uncommon accounting for only 2.5% of cases.^{9, 10, 11} Systemic risk factors for ION include cardiovascular disease, diabetes, and hypertension.¹²

The ocular pharmacology of SSRIs and their potential vascular side effects have been published.^{4, 5} Costagliola et al described a mechanism for vasospasm in the optic nerve, postulating that increased plasma serotonin levels may be a factor, or a cofactor, in the development of optic nerve perfusion disorders. With long-term SSRI treatment it was suggested that multiple transient vasospasms could progressively induce a manifest ischaemic optic neuropathy. In atherosclerotic individuals, the susceptibility to this disorder becomes higher due to serotonin-enhancing platelet aggregation on the atheromata of ocular arteries.^{4, 5, 9}

Recovery of vision is rare following nonvasculitic ischaemic optic neuropathy and therefore the question of a toxic aetiology remains open. These patients had been taking SSRIs for a mean of 7 years (range 1–14 years). These side effects may result from cumulative risk of prolonged treatment. In our series, we have identified smoking and diabetes as potentially synergistic risk factors in 4/5 cases.

Awareness of these potentially blinding vascular side effects is important. We advise caution when prescribing SSRIs in conjunction with known systemic vascular risk factors or pre-existing vascular eye disease until the origin of this neuropathy becomes clearer.

The authors declare no conflict of interest.