The contribution of integrons to the dissemination of extended-spectrum beta-lactamases (ESBL) was analyzed on all ESBL-producing Escherichia coli isolates from 1988 to 2000 at Ramon y Cajal Hospital. We studied 133 E. coli pulsed-field gel electrophoresis types: (i) 52 ESBL-producing clinical strains (C-ESBL) (16 TEM, 9 SHV, 21 CTX-M-9, 1 CTX-M-14, and 5 CTX-M-10); (ii) 43 non-ESBL blood clinical strains (C-nESBL); and (iii) 38 non-ESBL fecal isolates from healthy volunteers (V-nESBL). Class 1 integrons were more common among C-ESBL (67%) than among C-nESBL (40%) or V-nESBL (26%) (P < 0.001) due to the high number of strains with bla(CTX-M-9), which is linked to an In6-like class 1 integron. Without this bias, class 1 integron occurrence would be similar in C-ESBL and C-nESBL groups (47% versus 40%). Occurrence of class 2 integrons was similar among clinical and community isolates (13 to 18%). No isolates contained class 3 integrons. The relatively low rate of class 1 integrons within transferable elements carrying bla(TEM) (23%) or bla(SHV) (33%) and the absence of class 2 integrons in all ESBL transconjugants mirror the assembly of translocative pieces containing bla(TEM) or bla(SHV) on local available transferable elements lacking integrons. The low diversity of class 1 integrons (seven types recovered in all groups) might indicate a wide dissemination of specific genetic elements in which they are located. In our environment, the spread of genetic elements encoding ESBL has no major impact on the dispersion of integrons, nor do integrons have a major impact on the spread of ESBL, except when bla(ESBL) genes are within an integron platform such as bla(CTX-M-9).