Background: Due to concerns about hyperkalemia, the recommendation for patients with chronic kidney disease (CKD) is to limit intakes of foods high in potassium, including certain types of nuts, fruits and vegetables (F&V). Detailed patterns of F&V intake have not been described in patients with CKD, limiting our ability to study higher or lower risk patterns. In this study, we aimed to characterize the patterns of F&V intake in adults with and without CKD in a nationally representative sample of the US. Methods: We included 16,183 adults, with (n= 3,225) and without (n= 12,958) CKD based on eGFR and albuminuria from the Third National Health and Nutrition Examination Survey (NHANES III). We calculated counts of different types of F&V consumed by the participants according to a 24-h recall interview reported in the NHANES III Individual Food File. Based on their phytonutrient and starch content, F&V were then categorized into: rich in polyphenols; rich in carotenoids; rich in glucosinolates, and high in starch. We also categorized consumption of legumes and nuts. Patterns of legumes, nuts, and F&V intake were identified using latent class analysis (LCA; LCCA package, R 3.0.1). We evaluated differences in serum carotenoids, vitamins A, E, and C as objective biomarkers to help validate patterns. Multinomial logistic regression incorporating survey weights was used to adjust association between CKD status and F&V patterns (SAS Institute, V.9.4). Results: LCA analysis classified the food consumption into 3 distinct patterns: low F&V/low starch; moderate F&V/high starch; and high F&V/moderate starch. Consumption of legumes and nuts was low in all three patterns. Pattern of higher F&V consumption had higher serum levels of carotenoids, vitamins A, E, and C (p for difference across patterns <0.0001). Unadjusted patterns of consumption were not different in patients with vs. without CKD (p=0.654). After adjustment for ethnicity, gender, body mass index, waist circumference, diabetes, and hypertension, patients with CKD were more likely to consume moderate F&V/ high starch (OR=1.23, p=0.0316) or low F&V/starch (OR=1.34, p=0.0001) compared to patients without CKD. Conclusions: Utilizing the LCA analysis and regression, we found higher likelihood of consuming low F&V patterns in patients with vs. without CKD in the US. Further studies are needed to evaluate outcomes to understand risks of benefits of F&V patterns for patients with CKD.