Late posterior segment relapses in a series of Vogt–Koyanagi–Harada disease

We recently reported that 73.8% of subjects diagnosed with Vogt–Koyanaga–Harada (VKH) disease experienced recurrences (Errera et al. 2011). Based on our experience, we believed that finding an exudative retinal detachment (ERD) that responded to corticosteroids or other medical treatment more than 1 year after onset is low. To better evaluate late posterior recurrences, we reviewed the records of a large series of 45 patients with VKH disease from four tertiary centres. The median follow-up duration was 34 months since initial presentation. Late posterior segment relapses were defined as posterior segment manifestations of inflammation occurring later than 52 weeks after first presentation of VKH (Garcia-Valenzuela et al. 2000; Sachdev et al. 2008; Dolz-Marco et al. 2011). The study adhered to the tenets of the Declaration of Helsinki. Relapsing ERDs were noted within 10 months after presentation in 16 (84%) of 19 subjects who experienced such reactivation, with only four experiencing late ERDs, one of whom also had early relapses. Report of four subjects with late ERDs relapses at a median of 19 months (16–25 months). By examination and optical coherence tomography (OCT), subretinal fluid was limited localized areas of submacular ERDs in all late posterior segment relapses (Fig. 1). Intravenous pulse therapy was used within the first month for three of the four subjects who experienced a late posterior reactivation. In the fourth subject, the diagnosis of VKH disease was made 3 months after the first inflammatory signs, which caused a delay in the initiation of the correct treatment. This latter subject had treatment with corticosteroids for only 1 month, while the other three subjects received an initial duration of systemic corticosteroids for at least 6 months. In all four subjects, immunosuppressive therapy (infliximab, cyclosporin or interferon) was initiated after the first relapse because response to therapy with corticosteroid at acceptable long-term doses was insufficient. All four subjects had been on immunomodulatory agents (cyclophosphamide, infliximab and azathioprine), but these agents had been stopped 5 to 13 months prior to the relapse. In one subject, the relapse occurred while he was taking 25 mg/day of prednisone. He had responded to reintroduction of high-dose corticotherapy for an episode 6 months previously, and he was on tapering of therapy. The second subject was receiving both prednisone and cyclophosphamide. The third was on infliximab and low-dose oral corticosteroids at the time of relapse. In the experience at our institutions there were 4 late relapses (ERDs) of 45 subjects with an incidence rate of 12 cases per 1000 VKH affected person-years. Late relapses as ERDs occurred in patients receiving early intravenous pulse corticosteroids, and all four had been on other corticosteroid-sparing immunomodulatory therapy at some point. One possible explanation is that the need for more aggressive treatment in our patients with late relapses of VKH may be a marker of more severe disease. Early corticosteroid-sparing immunomodulatory therapy might be useful as first-line therapy to reduce the development of complications (Abu El-Asrar et al. 2013). Although it is difficult to conclude because we do not have enough statistical power (tests of Kaplan-Meier), we observed that subjects who had a longer duration of high dose corticosteroids (>3 months) at presentation of the disease were more likely to present posterior segment reactivation than those with shorter duration of corticosteroids therapy (This comparison raises the issue of resistance to corticosteroids treatment (and/or resistance to immunosuppressive therapies), of various therapeutics. We found, however, that late ERDs are limited to local submacular ERDs in our patients; larger ERDs do seem to be very rare. And if found, alternative diagnoses such as idiopathic chorioretinopathy (CSC) or rhegmatogenous detachment need to be considered. Conversely to what has been reported in CSC, no pigment epithelial detachment was seen on OCT in our cases. In all four subjects, ERD resolved following an increase or reinitiation of corticosteroids therapy which is another argument for excluding CSC. ‘Chronic’ or ‘recurrences’ are terms referring to subjects being off treatment or having discontinued the treatment (SUN terminology); however, in our series, among patients experiencing an ERD relapse, few of them may still be on treatment. We used the term ‘relapse’ instead.