There were three objectives to the present study: (1) compare the bladder infection rate and extent of biofilm formation for seven untreated spinal cord injured (SCI) patients and seven given prophylactic co-trimoxazole, (2) identify a level of bacterial adhesion to bladder cells which could be used to help predict symptomatic infection, and (3) determine from in vivo and in vitro studies whether fluoroquinolones were effective at penetrating bacterial biofilms. The results showed that the infection rate had not changed with the introduction of prophylaxis. However, the uropathogenic population had altered subsequent to the introduction of prophylaxis with E. coli being replaced by E. faecalis as the most common cause of infection. In 63% of the specimens from asymptomatic patients, the bacterial counts per cell were <20, while 81% of specimens from patients with at least one sign and one symptom of urinary tract infection (UTI) had > 20 adherent bacteria per bladder cell. Therefore, it is proposed that counts of > 20 bacteria adherent to sediment transitional epithelial bladder cells may be predictive of symptomatic UTI. Clinical data showed that fluoroquinolone therapy reduced the adhesion counts to <20 per cell in 63% of cases, while trimethoprim-sulfamethoxazole only did so in 44%. Further in vitro testing showed that ciprofloxacin (0.1, 0.5 and 1.0 micrograms/ml) partially or completely eradicated adherent biofilms from 92% of spinal cord injured patients' bladder cells, while ofloxacin did so in 71% cases and norfloxacin in 56%. These findings have important implications for the detection and treatment of bacteriuria in spinal cord injured patients.