The diagnosis of catheter-related infections relies on the presence of clinical manifestations of infection and the evidence of colonization of the catheter tip by bacteria, mycobacteria, or fungi. The reference method to confirm the latter requires the withdrawal of the catheter for culturing, which frequently turns out to be inconvenient, unnecessary and costly. New methods try to avoid these drawbacks and to assess the presence of tip colonization without withdrawal. Comparative quantitative blood cultures with a marked increase (> or = 5) in colony counts between blood obtained from the catheter lumen and from a peripheral vein simultaneously is one of those methods. It has a high sensitivity (>80%) and specificity (94-100%) but it is cumbersome and requires both an easy backflow of blood in the catheter and the existence of bacteremia. Cytocentrifugation and acridine orange staining of blood withdrawn from an infected catheter lumen has a sensitivity and a specificity of over 90% for the diagnosis of tip colonization. 'Superficial cultures' comprise the semiquantitative culture of the hub, of the skin surrounding the catheter entrance and of the first subcutaneous portion (1 cm) of the catheter after swabbing. The sensitivity of this method is >90%, specificity is >80%, and positive and negative predictive values for catheters (considering together those with and without clinical data of infection) are 66 and 97%, respectively. Endoluminal brushing has proved to be an impractical and unreliable procedure, at least in our experience. New methods based on the speed of bacterial growth to detectable levels of micro-organisms in conventional blood cultures are a new and interesting way of assessing catheter-related infections. Moreover, as the use of antimicrobial-coated catheters becomes more prevalent, the existing definitions of catheter colonization and catheter-related infection may need to be modified, because such coatings may lead to false-negative culture results. Many catheter infections, diagnosed without catheter withdrawal, can be handled nowadays with the so-called 'antibiotic lock-in technique', which consists in 'locking' the infected catheter lumen with a solution containing antibiotics. A high proportion of infected catheters, mainly those with coagulase-negative staphylococci, can be maintained in place and sterilized with this technique, including catheters in patients with therapeutic failure after receiving conventional intravenous antibiotic therapy. New diagnostic and therapeutic techniques may avoid the unnecessary withdrawal of thousands of efficient, difficult to replace and expensive intravascular lines.