Outcomes of patients with acute Vogt–Koyanagi–Harada disease treated with intravenous corticosteroid pulse followed by the slow tapering of oral corticosteroid therapy

To elucidate the intravenous corticosteroid pulse treatment outcomes of patients with acute Vogt–Koyanagi–Harada (VKH) disease and assess the differences between patients with no inflammation worsening and those with persistent or worsening inflammation. Potential factors responsible for eyes with low visual outcomes were also investigated. We retrospectively reviewed the clinical records of patients with acute VKH disease who first visited us between 2009 and 2018 and were followed up for > 300 days. Clinical characteristics, treatments, and posttreatment conditions were assessed. Patients were classified into no inflammation worsening (acute–resolved [AR]) and inflammation worsening (chronic–recurrent [CR]) groups based on conditions after 6 months from disease onset. This study included 60 eyes from 30 patients (mean age: 52.7 years). Patients were treated with methylprednisolone pulse followed by the slow tapering of oral prednisolone; 73% of patients developed AR and 27% CR, and the best-corrected visual acuity (BCVA) was ≥ 1.0 in 83% of eyes at 6 months following the introduction of treatment. Although the total prednisolone dose was higher in patients with CR disease, no significant difference was noted in the final BCVA. Among the patients, five eyes had a final BCVA of ≤ 0.5 due to anisometropic amblyopia, diabetic maculopathy, pre-existing macular hole, epiretinal membrane, and ellipsoid zone loss. Patients with acute VKH disease treated with corticosteroid pulse appear to demonstrate good visual outcomes, including patients with CR; the majority of eyes with low visual outcomes have pre-existing conditions that explain the low vision.

uveitis with choroiditis representing choroidal thickening, serous retinal detachments, and iritis [2]. Although appropriate treatment enables the majority of patients to exhibit monophase ocular symptoms, a small number of patients may develop recurrent or chronic inflammation, which may result in worse visual outcomes. Early treatment with high-dose corticosteroids (oral or intravenous) has been reported to reduce both the recurrence rate and related chronic inflammation [4][5][6][7][8][9][10]. Nakayama et al. reported that aggressive corticosteroid treatment with pulse intravenous corticosteroids followed by the slow tapering of oral corticosteroids in acute VKH disease resulted in excellent visual outcomes and low rates of recurrence [11]. Even with their aggressive corticosteroid treatment, 22.5% of patients had recurrent inflammation and some patients had low visual outcomes. Sakata et al. divided patients into acute-resolved (AR) with no inflammation worsening, chronic-recurrent (CR) with persistent or worsening inflammation, and CR with subretinal fibrosis (SRF) groups based on the worsening or persistent inflammation after 6 months from disease onset [12]. Consequently, the authors distinguished into periods where the initiation of treatment can influence the respective outcomes.
This study aimed to investigate the outcomes of acute VKH disease treated primarily with corticosteroid pulse, followed by the slow tapering of prednisolone, and assess the difference in clinical characteristics and visual outcomes between patients with AR and CR diseases. In addition, we investigated potential factors responsible for eyes with low visual outcomes.

Subjects
We retrospectively reviewed the clinical records of consecutive patients with VKH disease who first visited the Uveitis Division of Ophthalmology Department, Asahikawa Medical University Hospital, Japan, between January 2009 and March 2018. We included patients with complete, incomplete, and probable VKH based on the revised diagnostic criteria for VKH disease (2001) [13]. We excluded chronic and recurrent patients at their first visit. Thirty-one patients were able to follow the treatment course for > 300 days, and one patient who received topical therapy alone due to systemic condition of alcoholrelated hepatic insufficiency was excluded. A total of 30 patients (10 men and 20 women) were included for the main analysis of this study (Fig. 1). For reference, we also analyzed 31 patients including one who did not receive systemic corticosteroid.
All procedures performed in this study were in accordance with the 1964 Helsinki Declaration and its later amendments. This retrospective study was approved by the institutional review board of Asahikawa Medical University, Japan (approval number: 18257), and the requirement to obtain informed consent was waived.

Treatment
In general, our usual regime to treat acute VKH disease begins with corticosteroid pulse therapy (intravenous methylprednisolone 1,000 mg daily for 3 days), followed by the prescription of oral prednisolone 1 mg/kg/body weight/day. In certain cases, a pulse with intravenous methylprednisolone 500 mg daily for 3 days or a subtenon triamcinolone acetonide injection (STTA) would be selected instead of a 1,000 mg methylprednisolone pulse therapy at the discretion of the attending ophthalmologist. In cases where inflammation does not recur, the dose of prednisolone would be tapered to 20 mg after 2 months from the beginning of the corticosteroid pulse therapy and terminated after 6-9 months from the beginning of the treatment. We usually follow hospitalized patients prescribed with > 30 mg oral prednisolone/ day. Betamethasone 0.1% eye drops are prescribed to patients who had iritis, whereas mydriatic drops are prescribed to patients who have posterior synechia or whose iritis can cause posterior synechia. Since 2012, we have been performing optical coherence tomography (OCT) on each outpatient and weekly in hospitalized patients to monitor choroidal thickness and the presence of serous retinal detachment. Before 2012, we used to perform OCT only at the patients' first visit and when it was considered necessary to monitor serous retinal detachment. If serous retinal detachment and choroidal thickening do not worsen, we reduce the dose of oral prednisolone. In contrast, in cases where the worsening or recurrence of serous retinal detachment and choroidal thickening occurs, we consider an additional corticosteroid pulse therapy in patients receiving high-dose prednisolone. Furthermore, we either increase the prednisolone dose or prescribe a combination of cyclosporine in patients who have not been receiving high-dose prednisolone. Because cyclosporine use for noninfectious uveitis had not been covered by health insurance services until 2013 in Japan, we were unable to prescribe this drug before 2013. When encountering patients with mild choroidal thickening without undulations of the retinal pigment epithelium and no serous retinal detachment, we were initially trying to maintain the dose of prednisolone to observe whether patients' condition would be prolonged or whether they would recover. In cases where the patients' condition was prolonged during their next visit after approximately 1-2 weeks, we increased the dose of prednisolone in combination with or without cyclosporine. Finally, in cases where we observed recurrent iritis and no signs of the disease in the posterior segments, we increased the frequency of betamethasone 0.1% eye drops.
We usually co-prescribe a proton pump inhibitor or H 2 blocker but also amphotericin B oral suspension with systemic corticosteroid, provided that patients have no contraindications. To prevent glucocorticoid-induced osteoporosis, we prescribe alendronate sodium hydrate and alfacalcidol as the first and second line of defense, respectively, based on the guideline provided by the Japanese Society for Bone and Mineral Research [14]. Moreover, we prescribe sulfamethoxazole-trimethoprim to prevent pneumocystis pneumonia in certain patients who do not suffer from asthma, renal or liver dysfunctions, or other contraindications.

Data collection and analysis
In this study, we retrospectively reviewed patients' medical records to collect data regarding their sex, age at onset, site of inflammation, presence of extraocular symptoms, best-corrected visual acuity (BCVA), presence of HLA-DR4, HLA-DRB1*04:05, treatment features (the period from the onset to the beginning of systemic corticosteroid treatment, the period from the first visit to the beginning of the treatment, and the period of corticosteroid administration), and patient's conditions after the treatment (the presence of sunset-like fundus, BCVA, or complications).

Statistical analysis
For statistical analysis, we divided the patient course into AR with no inflammatory worsening, CR with persistent or worsening inflammation, and CR with SRF based on the presence of disease worsening or the persistence of anterior chamber cells after 6 months from the onset in accordance with the reports of Sakata et al. [12]. Based on the Standardization of Uveitis Nomenclature guidelines [15], we defined worsening inflammation as a two-step increase in the level of inflammation. Furthermore, we considered the conditions as "worsening inflammation" when the additional dose of prednisolone was necessary for the recurrence of serous retinal detachment or apparent choroidal thickening observed on OCT. We statistically compared the clinical characteristics between the patients with AR and CR diseases using Fisher's exact and Wilcoxon rank sum tests. For these analyses, the R version 3.6.3 (The R Foundation for Statistical Computing Platform, Vienna, Austria) and EZR (Jichi Medical University, Saitama, Japan) [16], a graphical user interface for R, were used.
For reference, we examined the number of patients who developed "recurrent" and "chronic" courses throughout the study period based on the definition of uveitis descriptors in the Standardization of Uveitis Nomenclature guidelines: "chronic" course as persistent uveitis with relapse in < 3 months posttreatment discontinuation, and "recurrent" course as the repeated episodes of uveitis separated by periods of inactivity without treatment for at least 3 months (Supplementary Information S1). We also statistically compared the clinical characteristics between patients who developed "recurrent/chronic" course and patients who did not (Supplementary Information S1, S2, S3). In addition, we statistically compared complete and incomplete VKH patients with probable VKH patients (S4).

Treatment
Among 31 acute VKH patients, we treated 30 patients with systemic corticosteroid therapy and they were included in the present study (Fig. 1). More specifically, 28 patients received 1,000 mg of intravenous methylprednisolone daily for 3 days, and 2 patients initially received 500 mg of intravenous methylprednisolone daily for 3 days. Treatment with 500 mg, instead of 1000 mg, methylprednisolone daily was selected as one patient was 80 years old and the other had a previous history of a cerebral hemorrhage. Two courses of corticosteroid pulse therapy were performed on one patient who had increased serous retinal detachments when prescribed 40 mg prednisolone. For this patient, the first course of corticosteroid pulse was 1,000 mg methylprednisolone daily for 3 days, and the second course was 500 mg methylprednisolone daily for 3 days.
The period from the onset of ocular symptoms to the beginning of the corticosteroid pulse therapy was between 4 and 63 (mean ± standard deviation: 18 ± 13) days and from the first visit to the start of corticosteroid pulse treatment was between 0 and 31 (mean ± standard deviation: 4 ± 7) days. Systemic corticosteroid treatment was completely tapered off in 28 patients during the study period and was continued beyond the study period in two patients. The mean total period of corticosteroid usage in patients who terminated their systemic corticosteroid treatment was 261 days, and the mean total period of corticosteroid usage in all patients, including those who are still undergoing systemic treatment, was 275 (range, 185 -508) days. Eight (27%) patients required increased corticosteroid doses because of serous retinal detachment recurrence or choroidal thickening, two of whom were also prescribed cyclosporine during the follow-up period (Table 1).

Course and visual outcomes
In total, 12 (40%) patients developed sunset glow fundus (Table 1), whereas no patients developed severe SRF. Figure 2 presents the changes in BCVA. BCVA was ≤ 0.5 in 37% of the eyes before treatment. In contrast, it was ≥ 1.0 in 65% of eyes at 1 month, 81% at 3 months, and 83% at 6 months following the introduction of systemic corticosteroids and was 78% at patients' last visit in the follow-up period. Although the final BCVA was ≤ 0.5 in five eyes (8%), the reasons for poor BCVA in three patients were considered to be anisometropic amblyopia, diabetic maculopathy, and the history of pars plana vitrectomy (PPV) for a pre-existing macular hole. Consequently, we did not relate these causes to VKH. However, the reason for the low final BCVA of ≤ 0.5 in the other two eyes was possibly related to the current disease. More specifically, we identified that VKH exacerbated retinal damage for one eye (BCVA: 0.5) with pre-existing epiretinal membrane (ERM) construction and that VKH induced ellipsoid zone loss in the macular of one eye (BCVA: 0.05; Supporting information S5 Figure). The mean final logMAR BCVA was -0.042 in the better-seeing eye and 0.099 in the other eye (Table 1). In additional analysis (Supplementary Information S1), although "chronic" course was observed in seven patients (23%), none of the 30 patients showed "recurrent" course. Results of statistical comparison of patients who developed "chronic" course and patients who did not develop "chronic" course are shown in Supplementary Information S2 and S3. Although sunset glow fundus was developed in significantly more patients who developed chronic course than those who did not on the analysis with 31 patients including one who did not receive systemic corticosteroid (p = 0.04), there was no significant difference in the analysis with 30 patients who received systemic corticosteroid. Other results were similar to those of the AR and CR comparisons. In the statistical comparison of complete and incomplete VKH patients with probable VKH patients, the only significant difference was observed in presence of extraocular symptoms (S4).

Discussion
In this study, we retrospectively analyzed the clinical characteristics and visual outcomes of 30 patients (60 eyes) with acute VKH disease treated with methylprednisolone pulse followed by prednisolone tapering. Five (8%) eyes had a final BCVA of ≤ 0.5, out of which for three eyes the reasons for low visual outcome were other than the current disease. Comparing the patients with AR and CR diseases who received systemic steroid therapy based on the inflammation condition at 6 months after onset, the total prednisolone dose was higher and the period of prednisolone use was longer in patients with CR disease. Final BCVA was good in both groups and there were no significant differences. The mean age of study participants was approximately 10-20 years older than that of the patients included in the majority of other studies [9,11,12,[18][19][20][21][22]. This may be attributed to the population age structure in northern Japan where our hospital is located. The remaining patient-related characters were almost similar to the ones included in previous studies. The proportion of patients with complete VKH disease was low in this study, similar to that in most previous studies [9,11,23,24]. This low proportion was speculated to be due to the fact that integumentary findings (alopecia, poliosis, or vitiligo) do not precede the onset of ocular symptoms, which are required to diagnose complete VKH, and integumentary findings may not always develop. In this study, HLA DR4 was positive in 90% and HLA DRB1*0405 was positive in 78% of patients, respectively. Of Japanese patients with VKH disease, > 90% have been reported to have HLA DR4, whereas approximately 24% of Japanese individuals have HLA DR4 and 13% have the HLA DRB1*0405 allele [11,25,26].
Early treatment with high-dose corticosteroids with the slow tapering of the therapy (for > 6 months) was reported to reduce the recurrence/prolongation of inflammation that resulted in good visual outcomes [4-7, 9, 20, 22]. Because the beginning of the corticosteroid pulse therapy at an average of 4 days after the patients' first visit (18 days after onset), which is followed by the slow tapering of oral corticosteroid for > 184 days, the treatment plan used in this study must be considered as early treatment with adequate corticosteroid dose and tapering periods. Although no differences in visual prognosis between intravenous corticosteroid pulse therapy and oral administration in acute VKH disease were reported [27], highdose intravenous pulse corticosteroid was reported to induce less choroidal change during the convalescent stage compared with oral corticosteroids [28]. Good visual outcomes have also been reported in patients with acute VKH disease treated with corticosteroid pulse therapy [11,22], and the visual outcomes of the present study are considerably better than other studies [18,19,28]. Starting treatment with corticosteroid pulse therapy can eliminate VKH inflammation more quickly than starting treatment with oral corticosteroids, and possibly lead to a better visual prognosis. The benefits of starting corticosteroid pulse therapy, instead of high-dose oral corticosteroid, include the rapid resolution of serous retinal detachment and choroidal thickening; simultaneously, this approach makes it easier to assess the effect and potential inflammation recurrence after the beginning of the treatment. Nakayama et al. treated patients with VKH disease with aggressive corticosteroid pulse therapy and achieved excellent visual outcomes [11]. The authors provided two and three pulses in 44% and 5.4% of their patients, respectively, based on the findings of fundus angiography performed 5-7 days after the last day of intravenous corticosteroid pulse therapy. In contrast, we gave a second course of corticosteroids to only one patient among the 30 patients (3%). We usually perform fundus angiography only on the patients' first visit to diagnose VKH and only provide a second course of corticosteroid pulse to patients whose choroidal thickening and serous retinal detachment recur while receiving high-dose prednisolone such as 40 mg. BCVA of ≥ 1.0 was noted in 86% of the eyes included in their study compared with the 65% in the present study at 1 month after the beginning of the systemic corticosteroid treatment. Approximately 94% of the eyes were included in the study by Nakayama et al. compared with the 78% in our present study at the final examination of the study period. However, the prevalence of developing sunset glow fundus in our patients was not higher than that in the study by Nakayama et al. (40% in the present study and 49.5% in the study by Nakayama et al.). This may indicate that the rate of patients conserved to be healed before depigmentation occurs in the fundus [29] is similar to that of their study [11]. Because most eyes with low visual outcomes in the present study had a pre-existing condition, such as ERM and macular hole, and the prevalence of these types of diseases is higher in older generations, the older age of patients in the present study may be a prevalent reason that facilitates lower visual outcomes than the study by Nakayama et al. When comparing systemic corticosteroid treatment side effects, no significant difference was identified and serious adverse effects such as sepsis and heart failure [30] were not observed in either study. Still, the possibility of aggressive corticosteroid pulse therapy, including multiple courses of corticosteroid pulses, contributing to the rapid visual recovery and better final BCVA, suggests that further comparison studies regarding the side effects must be conducted.
We classified patients based on AR and CR according to the study of Sakata et al. [12], assessing inflammation at 6 months after onset. According to the study by Sakata et al., 21% of patients developed AR, and 79% developed CR (41% developed CR with SRF), whereas the rates in our study were 73% for AR and 27% for CR and none had developed SRF. There are differences between these two studies in the patients' ethnicity, age (mean patient age in our study was approximately 20 years higher than theirs), and the treatment method followed. In the study by Sakata et al., the treatment started with oral prednisolone 1-1.5 mg/kg/day or methylprednisolone pulse therapy, followed by oral prednisolone with slow tapering over at least 6 months; they did not report how many patients had received methylprednisolone pulse. Although the potential ethnical differences and younger age of the patients in the study by Sakata et al. may be related to the recurrence rate, beginning treatment with corticosteroid pulse therapy and rapidly resolving inflammation may have resulted in the lower recurrence rate identified in the present study.
Sunset glow fundus was observed in 63% of patients with CR disease and in 32% of those with AR disease in the present study; however, there was no statistical difference (p = 0.21) even analyzed with the 31 patients including one who did not receive systemic corticosteroid (p = 0.11, Supplementary  Information S6). Conversely, sunset glow fundus was observed in more patients who developed a "chronic" course than those who did not develop a "chronic" course analyzed with patients including a patient who did not receive systemic corticosteroid (p = 0.04, Supplementary Information S6). Chronic course was defined as uveitis with relapse in < 3 months after treatment discontinuation, whereas CR and AR were classified based on the condition of inflammation at 6 months after onset when treatments were still ongoing. This difference in the definition and the small sample size might have influenced the results. The corticosteroid period was significantly longer and the total dose of prednisolone was significantly higher in patients with CR disease than in those with AR disease; however, no other differences were found regarding the clinical characteristics and outcomes such as age, sex, and final BCVA. In the present study, the visual outcome was good even in patients with CR disease and there was no significant difference in the final visual acuity between the patients with AR and CR diseases. This must have resulted from the fact that no patients developed severe SRF or retinal damage, not even in the CR group, in the present study.
Our study has several limitations. First, our design involves a retrospective study, and some of our findings such as skin findings might have been overlooked. Second, our sample size was relatively small and an increased number of patients may lead to different results, for example, significant differences between patients with AR and CR diseases in developing sunset glow fundus. Third, our patients were older than those in other studies and their ethnicity was limited to Japanese. Therefore, the effectiveness and the risk of developing adverse corticosteroid effects may differ according to the patients' background.
In conclusion, patients with acute VKH disease treated with corticosteroid pulse followed by the slow tapering of corticosteroids appear to demonstrate good visual outcomes without leaving SRF, even in patients with CR disease. The majority of eyes with low visual outcomes had pre-existing conditions that explained the low vision, such as ERM and macular hole.
Ethical approval All procedures performed in studies involving human participants were in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This study was approved by the institutional review board of Asahikawa Medical University (approval number: 18257).

Consent to participate
The requirement to obtain informed consent was waived by the institutional review board of Asahikawa Medical University.