Background/Aim: The possible association between lipoprotein(a) [Lp(a)] and albumin excretion rate (AER) is a topic that has generated conflicting views. The aim of this study was to determine the relationship between serum Lp(a) concentrations and AER in diabetic patients, taking into account Lp(a) phenotypes in a multivariate analysis. Methods: For this purpose 191 consecutive diabetic patients (69 type 1 and 122 type 2) were included in the study. Lp(a) was determined by ELISA and its phenotypes by SDS-PAGE followed by immunoblotting. Lp(a) phenotypes were grouped by size in small (F, B, S1, S2), big (S3, S4) and null. Results: Diabetic patients with an AER >20 µg/min presented higher Lp(a) concentrations than patients with an AER <20 µg/min: median 19 mg/dl versus 5 mg/dl (p < 0.0001). The differences remained at a significant level when the type of diabetes was considered. A linear correlation was observed between Lp(a) concentration and AER (type 1: r = 0.32, p = 0.01; type 2: r = 0.25, p < 0.05). The AER was independently correlated with Lp(a) concentrations in a multiple regression analysis (p < 0.01), and Lp(a) was independently associated with the presence of diabetic nephropathy in the logistic regression analysis. The overall frequency distribution of Lp(a) phenotypes differed significantly between patients with or without microalbuminuria (p < 0.05). In addition, the AER (µg/min) was different among the Lp(a) phenotypes: small 55 ± 122 (median 4.9), big 58 ± 123 (median 5.7) and null 3 ± 2 (median 2.3); p = 0.01. The significant difference mainly resulted from low AER (<10 µg/min) detected in all patients with the null phenotype. Conclusions: In diabetic patients the serum Lp(a) concentration is associated with AER. Thus, the elevated cardiovascular risk observed in diabetic patients with a high AER could be related to the Lp(a) concentration. Finally, patients with the null Lp(a) phenotype can be considered as a group at low risk of the development of diabetic nephropathy.