Retinal capillary nonperfusion seems crucial in the pathogenesis of sight-threatening diabetic retinopathy (DR); currently, no treatment prevents or reverts it. To further the understanding of the association between retinal capillary nonperfusion and sensitivity in DR. This prospective, longitudinal cohort study was conducted from April 18, 2018, to September 9, 2024, at a single center in the UK. Participants were followed up for up to 2 years; outcome assessors were masked. Adults (aged ≥18 years) with moderate or severe to very severe nonproliferative or proliferative DR with less than high-risk characteristics; at least 1 eye naive to treatment; no other retinal disorders; who were able to provide informed consent; and who were willing undergo retinal imaging were eligible for inclusion. Data analysis was performed from September 2024 to April 2025. The primary outcome was the association between retinal sensitivity (110° projection perimetry) and retinal perfusion (ultra-widefield angiography) at baseline and changes at 1 and 2 years in the study eye. Of 66 people approached, 50 were eligible and recruited, and 44 individuals with at least 1 perimetric examination were included. Mean (SD) participant age was 52.1 (12.2) years, and 13 participants (29%) were female. Median hemoglobin A1c was 75.5 mmol/mol (9.1% of total hemoglobin [to convert from percentage of total hemoglobin to proportion of total hemoglobin, multiply by 0.01]); mean (SD) best-corrected visual acuity letter score was 85.7 (4.7) (Snellen equivalent, 20/20). Mean retinal sensitivity deficit at baseline was associated with perfusion status, with larger deficits in nonperfused areas (n = 354; 11.8 dBs; 95% CI, 10.8-12.8) compared to perfused areas (n = 2092; 6.6 dB; 95% CI, 5.1-8.2; P < .001). Only age correlated positively with sensitivity deficit (estimate, 0.2; 95% CI, 0.1-0.3; P = .006). A deficit of 5 dB or greater occurred in 711 of 2092 (34%) perfused areas; 105 of 354 (30%) nonperfused areas had normal sensitivity. Rates of sensitivity deficit change in perfused and nonperfused areas from baseline to 1 year were -0.20 dB/mo (95% CI, -0.24 to -0.16) and -0.28 dB/mo (95% CI, -0.41 to -0.15) (perfused vs nonperfused, P = .22), respectively (1464 areas); from baseline to 2 years, rates were -0.16 dB/mo (95% CI, -0.20 to -0.12) and -0.34 dB/mo (95% CI, -0.47 to -0.21) (perfused vs nonperfused, P = .007), respectively (542 areas). In this longitudinal cohort study, although retinal capillary perfusion status was associated with function, sensitivity loss occurred in some perfused areas and normal function in some nonperfused areas; sensitivity deficit decreased over time (approximately 45% in the first year) despite poor glycemic control and high DR grades. These findings should be considered for the management of people with DR and the design of clinical trials.