In this study, we aimed to evaluate the association of asymptomatic optic nerve demyelinating lesion in patients presenting a clinically isolated syndrome with the asymptomatic retinal neuro-axonal loss previously reported at clinically isolated syndrome. We prospectively recruited 66 patients presenting a clinically isolated syndrome and 66 healthy control subjects matched according to age and gender. All patients underwent brain magnetic resonance imaging including 3D-double inversion recovery (DIR) sequence, optical coherence tomography examination and visual function evaluation, at 2.5-4.5 months after CIS. Evaluation criteria were presence and length of optic nerve DIR hypersignal, retinal layers (including ganglion cell inner plexiform layer and inner nuclear layer) thickness/volume, and low contrast monocular vision acuity (number of letters correctly identified). All clinically isolated syndrome eyes with past history of optic neuritis (CIS-ON) presented an optic nerve DIR hypersignal. We observed asymptomatic optic nerve DIR hypersignal in 22.2% of clinically isolated syndrome eyes without optic neuritis (CIS-NON). In comparison with healthy control, GCIPL volume (in mm3) was significantly lower in CIS-ON eyes [β (95% confidence interval, CI) = -0.121 (-0.168 to -0.074); P < 0.0001], and to a lesser extent in CIS-NON [β (95% CI) = -0.023 (-0.039 to -0.008); P = 0.004]. In comparison to healthy controls, eyes with asymptomatic optic nerve DIR hypersignal presented significantly lower macular ganglion cell inner plexiform layer volume [β (95% CI) = -0.043 (-0.068 to -0.019); P = 0.001], and eyes without did not [β (95% CI) = -0.016 (-0.034 to 0.003); P = 0.083]. Among CIS-NON, macular ganglion cell inner plexiform layer volume decrease was associated with asymptomatic optic nerve DIR hypersignal independently of optic radiations T2 lesions and primary visual cortex volumes (P = 0.012). Symptomatic optic nerve DIR hypersignal were significantly longer (13.8 ± 6.7 mm) than asymptomatic optic nerve hypersignal (10.0 ± 5.5 mm; P = 0.047). Length of optic nerve DIR hypersignal was significantly associated with thinner inner retinal layers (P ≤ 0.001), thicker inner nuclear layer (P = 0.017) and lower low contrast monocular vision acuity (P < 0.05). Compared to healthy control, low contrast monocular vision acuity was significantly lower in CIS-ON eyes (P < 0.0001) and CIS-NON eyes with (P = 0.03) or without asymptomatic optic nerve DIR hypersignal (P = 0.0005). Asymptomatic demyelinating optic nerve DIR hypersignal at the earliest clinical stage of multiple sclerosis is frequent and associated with asymptomatic retinal neuro-axonal loss reported at clinically isolated syndrome stage. Length of optic nerve DIR hypersignal is a biomarker of retinal neuro-axonal loss and visual disability at clinically isolated syndrome stage. Visual disability of clinically isolated syndrome eyes without clinical and subclinical optic nerve involvement might be due to missed optic nerve lesions on MRI. At the earliest clinical stage of multiple sclerosis, our results support considering optical coherence tomography as a window to the optic nerve rather than to the brain.