Vogt-Koyanagi-Harada disease: a retrospective and multicentric study of 41 patients

The Vogt-Koyanagi-Harada (VKH) disease is a bilateral granulomatous panuveitis with potential systemic involvements: neurological disorders (cerebrospinal fluid analysis shows pleiocytosis in about 80% of cases), otological disorders (hearing loss, dizziness (70%) and tinnitus (42%)) and dermatological disorders such as vitiligo, poliosis and alopecia (10 to 63%) [1]. The disease is mediated by Th1 lymphocytes targeting melanocytes [2]. The origin of this affection remains unknown, though many infectious triggers have been hypothesized [2‐4]. The association with HLA DR4/ HLA DRB1–04*05 has been reported in the Japanese population [5]. VKH disease mainly affects subjects from East and South East Asia, as well as Amerindian and Hispanic patients [1, 6]. Median age at diagnosis is around 40. Only few studies are available to estimate prevalence of ethnic characteristics in the European population [7‐10]. The aim of this study was to describe patients’ epidemiological characteristics in clinical French referral centers.

Forty-one patients (25 women and 16 men) were included. The main characteristics are reported in Table 1.

Mean age at diagnosis was 38.7 years (10–74) and the median of follow-up was 29.7 months (1–140). North African was the main represented ethnicity 58% (24/41) followed by South East Asian 49% (20/41), Caucasian 20% (8/41) and Hispanic 2% (1/41). Thirty patients were diagnosed at the acute early onset stage of the disease (73%). Nine patients had a complete presentation (22%), 19 an incomplete presentation (46%) and 13 had a possible syndrome (32%). All patients with a dermatological manifestation were diagnosed at the late stage of the disease. First entry in the medical care system was mainly by the Ophthalmology department (92%, 37/41) and patients were mostly followed by both Ophthalmology and Internal Medicine departments (61%, 25/41).

VKH disease is a rare disease and most of epidemiological data has been described for Asiatic, Hispanic, and Amerindian populations but not so for patients from Northern Africa. To our knowledge, our study is the second largest French cohort [8, 10, 17]. The low prevalence of the disease in Europe explains the low number of patients by center [18].

Apart from ethnicity, epidemiological data is the same as that of other populations’ [2, 6] with 60% of women and a mean age 38.7 at diagnosis. In our study the prevalence of the association with HLA DR4/ HLA DRB1–04*05 agrees with the literature data, the presence of this allele being associated with a higher risk of developing this disease.

As described by Lavezzo et al. and Ohno et al., the meningeal involvement is variable (63% of cases) like the clinical meningismus [1, 19]. In our study cerebral imaging did not find any significant sign, since its role is not defined in VKH disease. Although this investigation is normal in most cases, it seems relevant to exclude differential diagnoses. The otologic involvement in our population is the same as that of non-Hispanic populations’ [6]. This involvement is typical (75% of cases with a mean hearing loss of 30 dB) but seldom symptomatic [6]. While VKH disease could usually be confidently diagnosed by uveitis specialists based on ocular presentations, a search of systematical involvement, in particular an audiogram and/or a lumbar puncture, might be helpful for physicians who are less familiar with this disease. The prevalence of cutaneous involvement is closed to the non-Japanese populations’ [20]. Skin involvement prevails in pigmented populations from 10 to 63% [6]. Poliosis, vitiligo and alopecia are late manifestations of the disease. Most of our patients were diagnosed at an early stage, but a specific attention should be given to the extra ophthalmic manifestations at diagnosis of first ophthalmological manifestation and during follow-up to avoid a late-stage diagnosis. In our study, we did not notice any delay between the first manifestation of the disease and the ophthalmic involvement that led to the ophthalmic examination, but most of our patients presented the ophthalmic signs first (93%, 38/41).

Sasamoto et al. had a better outcome with high doses of systemic corticosteroid therapy [21]. Oral prednisone equivalent is usually prescribed initially at a dosage of 1 to 1.5 mg/kg/day with progressive tapering [17, 22]. In serious situations, methylprednisolone can be used at 1 g/day for 3 to 5 days [2]. The corticosteroid therapy duration is not codified, but in our study a few patients showed persistent inflammation signs at 6 months [21]. Treatment should be adapted to each patient depending on the clinical outcome. An early cessation is associated with an elevated rate of treatment failure [23]. Surprisingly, in our study treatment failures occurred from 1 to 48 months after the beginning of treatment, with a mean of 13 months, at a high level of corticosteroid of 15 mg a day, though the literature suggests a decrease of corticosteroid therapy with a minimal duration of 6 months [24]. Nevertheless, we believe that prognosis should have been different if immunosuppressive therapy had been initiated at first line [25, 26].

The sooner the diagnosis, the better the disease outcome (visual acuity ≥5/10 after treatment) [1]. This data was found in our cohort with a positive average profit by eye after treatment. An immunosuppressive treatment was introduced in 27% of cases as first line of treatment or in case of failure. Ciclosporine, Azathioprine, Cyclophosphamide, Methotrexate and Mycophenolate mofetil are active therapeutics in this disease [2, 27‐29]. A recent Mexican study did not find benefit to adding early immunosuppressive treatment in terms of final visual acuity or of development of visually significant complications [25]. In contrast, other studies found a preeminence of immunosuppressive agents (Mycophenolate mofetil, Ciclosporine A, Methotrexate, and Azathioprine) over corticosteroids alone, while others were more reserved [26, 30]. Finally, biological therapies are used in non-infectious uveitis including VKH disease [31] but the treatment length in VKH disease is still debated.

Patients diagnosed at early stage presented mainly unilateral loss of visual acuity (p = 0.01). They were treated more often with bolus of corticosteroid (p < 0.01) and second line agents (p < 0.01). Compared to patients who were diagnosed at the early stage of disease, patients diagnosed at the chronic stage of the disease were treated longer: 36 months instead of 20 (p < 0.01). Complications were also more frequent in this group.

As in other French series, our study included most patients from Maghreb among a multiethnic population [8, 10]. Our prognosis seems better than in North African studies, which are more homogeneous as for ethical origin. Alaoui et al. reported about 8 women: 5 were cured without any relapse and only two were still treated with corticosteroids. Khairallah et al. found a good outcome for 59% of patients maintaining a visual acuity of 20/40 or better [32]. Boutimzine et al. observed 87.5% of patients with a visual acuity of 5/10 or better after a mean follow-up of 6 years [33]. We failed to find any significant clinical difference in the Maghreb population, but for age and gender trends. However, we report here one of the largest case reports about Maghreb to our knowledge [32‐34] suggesting a higher prevalence in this population. Indeed, initial clinical presentation seems more severe in this population, especially the ophthalmic presentation, even if we did not find any clinical significance [35]. In our study the initial visual acuity was low, and this data was associated with a worse prognosis factor [36]. Surprisingly, as opposed to Read et al. who observed 51% of complication with 27% of glaucoma, or to Pandey et al. who observed a glaucoma cumulative incidence of 11.7%, we noticed a lower rate of relapse (44%) and complication like glaucoma (5%), which is known to be a severe one [37]. Larger multiethnic studies are needed to explore the severity of the disease according to ethnic origins.

Our study has several limitations. The first is related to the retrospective design of the study, namely missing data; the second limitation is due to non-harmonized diagnoses strategies within the 3 university Hospitals; the third one is due to the small sample size: HLA type was only performed on 10 patients. Complications were also noted for a few patients. Furthermore, we did not notice the necessary time to make a diagnosis after the beginning of symptoms and the necessary time to treat patients. Finally, the treatment was individualized due to the lack of guidelines concerning this disease at time of diagnosis. The high recurrence rate of 44% could be due to sample bias (close follow-up, tertiary centers). Nonetheless, it is the second largest French cohort reported to date [8, 10, 17].

VKH disease is a rare affection, more present in Asian, Hispanic, and Amerindian populations but in North African populations as well. In French studies patients originated from Maghreb represent a significant proportion, but all ethnicities could and should be involved. Accordingly, we should not limit our vision to patients’ ethnic origins, but diagnosis should be brought up particularly when ophthalmic manifestations are associated to neurological, otologic and/or dermatological signs.