In this retrospective, multi-centre, observational study, the impact of the initial management of Vogt-Koyanagi-Harada syndrome (VKHS) was assessed by reviewing the records of all patients diagnosed with VKHS between January 2001 and December 2020 in two French tertiary centres, with patients with less than 6 months of follow-up excluded. Fifty patients were included with a mean age at diagnosis of 41.4 years old and a median duration of follow-up of 29.8 months. Forty-six of the patients received prednisone after methylprednisolone, five patients received at least one associated immunosuppressive therapy (IST) within the first six months and 26 patients received IST during the entire follow-up period. Immunosuppressive therapy was added for steroid-sparing purposes, steroid side-effects or at the physician’s discretion. Treatment delay was set at 26 days using the ROC curve approach. Twenty-eight patients (56%) had at least one relapse during follow-up, at a median of 5.4 months from diagnosis. Multivariate analysis was done, demonstrating a significant association between relapse and delayed treatment (>26 days) (HR=3.69, CI95% 1.30–10.47, p=0.01), whereas there was no association between relapse and the number of corticosteroid pulses at initial management. An analysis showed patients of Afro-Caribbean origin were significantly more likely to relapse (HR=3.18, CI95% 1.26–8.03, p<0.01) and there was a non-significant increase in relapse in patients treated 26 days after symptom onset (HR=2.07, CI95% 0.98–4.38, p=0.05). Overall, this study has shown that early corticosteroid treatment within the first 26 days of symptoms can decrease the relapse rate significantly. However, this study is limited by its retrospective design and small cohort size and further prospective studies are required to investigate the effects of IST as first-line on relapse rate.