Study design: A retrospective study concerning urinary tract infections in spinal cord injury (SCI) patients. Objectives: To check whether the regular (1/week) urine cultures allow a more accurate treatment of urinary tract infections in SCI patients compared to empiric treatment. Setting: Ghent University Hospital, East-Flanders, Belgium. Methods: Group 1: 24 tetraplegic patients; group 2: 22 paraplegic patients; group 3: 28 other polytrauma patients as controls. These groups were chosen as catheterisation and other voiding methods differ according to the underlying pathology.Results: An average of four clinically significant episodes of bacteriuria were found for groups 1 and 2, while group 3 experienced very few urinary infections. The mean species turnover of the first two groups was 2. No statistically significant difference was found in antibiotic-resistance patterns of organisms isolated. Conclusion: Despite different catheterisation techniques in para-and tetraplegic patients, we conclude that: (1) the number of episodes of clinical significant nosocomial urinary infections is not different; (2) the mean species turnover is the same; (3) because of the species turnover, the value of regular urine cultures for 'documented' treatment of clinical relevant urinary infections seems to be limited. So urine culture could be performed less frequently or only when therapy becomes mandatory; (4) No oral antibiotic with superior activity was found: treatment is best started empirically (after sampling for urine culture) and adjusted to the resulting antibiotic sensitivity screening.