Diabetic retinopathy (DR) is a duration-dependent complication of diabetes (DM). Yet some people with DM do not develop DR despite long disease duration. We evaluated such a group in search of novel factors that might signal protection from DR, using a large cohort of Latinos with type 2 DM and readable retinal images in the GOLDR study (n=614). Participants were phenotyped and 7-field retinal images were evaluated using Airlie House criteria. We identified 90 participants with DM>10y without evidence for DR (NoDR). We compared this group of patients with another group more susceptible to DR with evidence for earlier onset DR, in DM <10y duration (EoDR, n=103). Duration of diabetes in NoDR was [x+SEM] 14.2+ 0.6y, and in EoDR, 4.3+ 2.9 y (p<0.001), a 10-y spread. We found that most of the typical DR-associated risk factors could not explain DR protection in NoDR, including age, sex, age at DM onset, systolic blood pressure (SBP), percent insulin users, duration of hypertension, fasting plasma glucose, A1C, urine albumin/creatinine ratio and estimate glomerular filtration rate; these parameters were not significantly different in the two groups. Protective factors that did emerge were female sex (p=0.02), lower diastolic BP 69.1+0.9 vs. 72.5+0.9 (p<0.01) and lower alcohol intake 3.1+0.8 vs. 7.8+2 de/w (14g drink equivalents/week; p=0.025). In a sensitivity analysis to determine whether sex accounted for the apparent effect of alcohol on DR, we evaluated the men in the study, who were more likely to be drinkers. Alcohol consumption was compared in men with DR who reported drinking alcohol (n=93) compared to men without DR who also reported drinking (n=53). Men without DR reported significantly less alcohol intake, 14.8+2.4 vs. 25.9 +3.3 de/w in those with DR (P<0.01), suggesting that a possible protective benefit of lower alcohol consumption observed in NoDR was not likely to be mediated by the presence of fewer men in that cohort. In summary, type 2 diabetic patients with no evidence of DR after 10y were more likely to be women, have a lower diastolic BP, and who imbibed less alcohol when compared with a more accelerated DR subgroup with <10yrs duration of DM. We conclude that in type 2 DM Latino patients, a focus on alcohol intake may be a useful management strategy in addition to traditional medication-based BP control and renal protection, as well as a pathophysiological pathway for DR worthy of investigation.