We surveyed the extensive literature regarding the development, therapy and prevention of catheter-associated urinary tract infections (UTIs). We systematically searched for meta-analyses of randomised controlled trials available in Medline giving preference to the Cochrane Central Register of Controlled Trials and also considered other relevant publications, rating them on the basis of their quality. The studies' recommendations, rated according to a modification of the US Department of Health and Human Services (1992), give a close-to-evidence-based guideline for all medical disciplines, with special emphasis on urology where catheter care is an important issue. The survey found that the urinary tract is the commonest source of nosocomial infection, particularly when the bladder is catheterised (IIa). Most catheter-associated UTIs are derived from the patient's own colonic flora (IIb) and the catheter predisposes to UTI in several ways. The most important risk factor for the development of catheter-associated bacteriuria is the duration of catheterisation (IIa). Most episodes of short-term catheter-associated bacteriuria are asymptomatic and are caused by a single organism (IIa). Further organisms tend to be acquired by patients catheterised for more than 30 days. The clinician should be aware of two priorities: the catheter system should remain closed and the duration of catheterisation should be minimal (A). While the catheter is in place, systemic antimicrobial treatment of asymptomatic catheter-associated bacteriuria is not recommended (A), except for some special cases. Routine urine culture in an asymptomatic catheterised patient is also not recommended (C) because treatment is in general not necessary. Antibiotic treatment is recommended only for symptomatic infection (B). Long-term antibiotic suppressive therapy is not effective (A). Antibiotic irrigation of the catheter and bladder is of no advantage (A). Routine urine cultures are not recommended if the catheter is draining properly (C). A minority of patients can be managed with the use of the non-return (flip) valve catheter, avoiding the closed drainage bag. Such patients may exchange the convenience of on-demand drainage with an increased risk of infection. Patients with urethral catheters in place for 10 years or more should be screened annually for bladder cancer (C). Clinicians should always consider alternatives to indwelling urethral catheters that are less prone to causing symptomatic infection. In appropriate patients, suprapubic catheters, condom drainage systems and intermittent catheterisation are each preferable to indwelling urethral catheterisation (B).