engleski

Rising problem of multidrug resistant gram negative bacteria causing bloodstream infections after liver transplantation: how should we handle the issue?

We have read with great interest the article by Bert et al. concerning microbial epidemiology of bloodstream infections (BSIs) after liver transplantation (LT).1 We would like to comment on this issue based on our results, which support the reemergence of Gram-negative (GN) bacteria as a leading cause of BSI after LT. The use of antibacterial and antifungal prophylaxis through the decades has likely contributed to the emergence of increasingly resistant pathogens, now necessitating alterations in the antimicrobial therapy. After introduction of selective digestive decontamination (SDD) against GN microorganisms in the mid 1990-s, Gram-positive (GP) bacteria emerged as major pathogens. However in 2004 Singh et al. and now Bert et al. reported reemergence of GN bacteremias.1, 2 We analyzed BSIs occurrence in 136 consecutive LT patients (mean age 49 years) during the first year after LT. The sudy period was from January 2005 to October 2008. We considered BSI to be significant according to the Center for Disease Control (CDC) criteria: a) the isolation of a microorganism other than skin flora in at least one culture with the presence of signs of infection or b) the isolation of a skin flora microorganism in at least two consecutive cultures associated with signs of infection.3 Blood cultures were incubated in a BacT/Alert automated system (BioMerieux, France). Isolates were identified using standard microbiology procedures and sensitivity testing was done by disk diffusion using CLSI standards and interpretation.4 Minimal inhibitory concentrations were done by E-test (AB Biodisk, Sweden). All patients received perioperative prophylaxis with piperacilin-tazobactam (5 days) and SDD (gentamycine and fluconazole, 3 days). Standard prophylaxis with trimethoprim sulfamethoxazole (9 months) for Pneumocystis jiroveci and valgancyclovir (3 months) for Cytomegalovirus was applied in all patients. Immunosuppression consisted of prednisone induction and maintenance regimen: prednisone (withdrawal afer 3-6 months), calcineurin inhibitors (tacrolimus or cyclosporine) and mycophenolate mofetil. A total of 47 BSIs were identified in 23 (16.9%) transplanted patients, which represents an incidence of 34.6 episodes per 100 transplanted patients. They occurred predominantly in the first trimester (median of 46 days post LT, range 0-248, SD 72 days). Polymicrobial bacteremias accounted for 12.8% of cases. Altogether 53 pathogens were isolated. GN bacilli, predominantly Klebsiella pneumoniae, Pseudomonas aeruginosa, and Escherichia coli were responsible for 58.5% of BSIs (Table 1). High incidence (42%) of resistant enterobacteriaceae producing extended spectrum beta-lactamases (ESBLs) was observed among BSI isolates (Table 2). Low sensitivity in enterobacteriaceae was recorded for piperacillin/tazobactam (63%), 3rd generation cephalosporins (58%), cefepim (58%), gentamicin (47%), amikacin (53%) and ciprofloxacin (21%). High proportion of P.aeruginosa isolates resistant to multiple antibiotics, including carbapenems (57%), was found in our series of patients. The most prevalent GP microorganisms isolated were enterococci and coagulase negative staphylococci (CoNS). CoNS were regarded as causative agents of BSI if they were isolated in two or more blood cultures and a catheter tip. In cases of CoNS all BSIs were catheter related as determined by clinical criteria. Staphylococcus aureus (MRSA) was responsible for only one BSI. Fungi accounted for a higher proportion (13.2%) of isolates compared to other studies in which antifungal prophylaxis was used5. Most fungemias were caused by Candida albicans (50%), whereas non-albicans species included Candida parapsilosis and Candida tropicalis which were fluconazole-susceptible. Deaths occured within 15 days as result of BSIs in 8 of the 47 episodes (17%) and were predominanty asocciated with ESBLs producing K.pneumoniae (3 cases), carbapeneme non-susceptible P.aeruginosa (3 cases) and Candida parapsilosis (2 cases). The only vancomycin-resistant isolate of E.faecium (VRE) resulted in fatal outcome. In 3 patients that died of the BSI multiple organisms were isolated from blood cultures. In conclusion, our data support the recently described reemergence of GN pathogens as leading agents of BSI after LT. It is particularly worrisome that a large proportion of GN bacteria causing BSI are multiply resistant to antibiotics, compromising not only the empirical therapy of sepsis in LT patients but also leaving, in some cases, few choices even when the causative agent is known. Carbapenem resistant non fermentative bacteria are becoming more prevalent in our unit and it may only be a matter of time when colistin will no longer be effective as a drug of last resort. Rigorous infection control practices should be applied in an attempt to stop the spread of such organisms once they appear on the ward. Not less important is the need to rationalize the use of antibiotics which are in many cases misused in the treatment of bacterial colonization of different sites or unnecessarily prolonged duration of prophylaxis. Surveillance and analysis of bacteria causing serious infections in defined subsets of patients is an essential basis for interventions aiming at improvement of treatment and prophylaxis of infections in these patients. In our case we will try to reduce the rate of multiply resistant GN bacteria by reducing the duration of surgical prophylaxis to 24-48h and switching to a more narrow spectrum antibiotic. Reinforcement of contact isolation and standard precautions is a process currently underway in our hospital.

bloodstream infections; liver transplantation; multidrug resistance

nije evidentirano