CHICAGO – A new bedside score based on easy-to-assess clinical factors, such as the presence of diabetes and admission to the intensive care unit, may help reduce delays in appropriate antibiotic therapy for bloodstream infections caused by carbapenem-resistant Enterobacteriaceae.
Researchers developed the score using data from patients who had bloodstream infections due to Enterobacteriaceae that produced extended-spectrum beta-lactamase (ESBL). In 9% of the patients, the pathogen was resistant to carbapenem.
Study results, reported in a poster session at the Interscience Conference on Antimicrobial Agents and Chemotherapy, showed that the combination of five simple clinical factors yielded a carbapenem-resistant Enterobacteriaceae (CRE) score having an area under the receiver operating characteristic curve of 0.81, which indicates that the scoring tool would be good for identifying carbapenem-resistant infections.
"We have been trying ... to predict whether [patients] should empirically stick with the carbapenem or use colistin."
The model is being expanded and validated, according to lead investigator Emily T. Martin, Ph.D., an assistant professor in the department of pharmacy practice at Wayne State University, Detroit. The hope is that in its final form, the model will help rapidly determine which patients with ESBL-producing Enterobacteriaceae infections can receive a carbapenem (the usual empiric therapy) and which need an alternate antibiotic. Some patients suspected of having ESBL-producing Enterobacteriaceae could have infections that are resistant to carbapenems, but we don’t know whether that’s true until the lab results come back, which can take up to 5 days," she explained in an interview.
Polymyxins are effective against carbapenem-resistant infections, but they are not appropriate for widespread empiric use because of concerns about nephrotoxicity and the emergence of resistance.
"At Detroit Medical Center, there’s been sort of movement toward using colistin (polymyxin E) if physicians are suspecting that the patient might have a carbapenem-resistant infection, but there is really no data" to guide the practice, she said. "So we have been trying to develop a model using characteristics of the patients to try to predict whether they should empirically stick with the carbapenem or use colistin."
Dr. Martin and her colleagues reviewed the charts of 182 patients admitted to the medical center between 2007 and 2010 who had bloodstream infections due to ESBL-producing Enterobacteriaceae and severe sepsis, septic shock, or multiorgan failure. In 16 patients, the pathogen was resistant to carbapenem.
A multiple regression model was used to incorporate five factors into a score having a maximum possible value of 51 points: neurologic disease (14 points), dependent functional status (7 points), diabetes mellitus (12 points), ICU admission at the time of infection (11 points), and antibiotic exposure in the past 3 months (7 points).
Using a cutoff of 32 points or higher to define high risk, the CRE score had a sensitivity of 81%, a specificity of 70%, a positive predictive value of 21%, and a negative predictive value of 97% for CRE infections, according to results reported at the conference, which was sponsored by the American Society for Microbiology.
The CRE score "has a really high negative predictive value, so basically, using these factors, we’re able to pretty well rule out that [an infection] is going to be requiring colistin," commented Dr. Martin. "But we are not able to identify which patients are the best candidates for colistin so far."
As for the future? "We are trying to get more data so we can improve the model, and then try to look at data from other hospitals too to try to make this as broad a model as possible," she said.
Dr. Martin reported that she had no relevant conflicts of interest.