Editor's note: This story was updated on Jan 29, 2018, with comments from lead study author Joppe van Duijn, MD.
Since the 1980s, the concept of rotating antibiotic treatments for patients with bacterial infections has been viewed as a potential strategy for reducing antibiotic resistance rates in hospitals. The idea behind the strategy is that withdrawing the use of an antibiotic in favor of another on a hospital-wide level will reduce the selective pressure for resistance so that the drug will remain effective.
This approach, called antibiotic cycling, involves using a specific antibiotic as a first-line therapy for all patients for a defined period, replacing that antibiotic with a drug of a different class but a similar spectrum of activity for the same duration, then repeating the cycle. More than two different antibiotics can be part of this rotation. The strategy has been likened to crop rotation.
Another strategy for reducing the prevalence of drug resistance, called antibiotic mixing, takes a less structured approach, with each consecutive patient on a ward being treated with an alternative class of antibiotic.
These approaches have been of considerable interest in intensive care units (ICUs), where patients are vulnerable to antibiotic-resistant gram-negative bacterial infections, antibiotic usage is high, and transmission of drug-resistant bacteria frequently occurs. But neither antibiotic cycling nor mixing has produced conclusive results in clinical studies, and it's unclear if one strategy would be better for reducing rates of antibiotic-resistant gram-negative bacteria.
Now, a team of European researchers writing in The Lancet Infectious Diseases has concluded that neither strategy is particularly effective.
Strategies produce similar results
In the multi-center, cluster-randomized crossover study, the researchers randomly assigned eight European ICUs to two 9-month intervention periods in which they could use one of three antibiotic groups (third- or fourth-generation cephalosporins, piperacillin-tazobactam, or carbapenems) for empirical treatment of patients with ICU-acquired gram-negative bacterial infections. The intervention periods were separated by a 1-month washout period.
In the cycling intervention, the preferred empirical treatment changed every 6 weeks; during mixing, it changed with every consecutive empirical treatment course. Treating physicians could deviate from the protocol in case of safety concerns and were allowed to use non-study antibiotics for combination therapy.
Over the almost 3-year study period (June 2011 through February 2014), three ICUs were assigned to mixing followed by cycling, and five were assigned to cycling followed by mixing. The primary end point was unit-wide prevalence of carriage with antibiotic-resistant, gram-negative bacteria—defined as Enterobacteriaceae with extended-spectrum beta-lactamase production or resistance to piperacillin-tazobactam and Acinetobacter spp and Pseudomonas aeruginosa with resistance to piperacillin-tazobactam or carbapenems—measured through monthly point-prevalence screening cultures.
Overall, 4,069 patients were admitted to the ICUs during the cycling periods and 4,707 were admitted during the mixing periods. Of these patients, 1,598 (745 during cycling and 853 during mixing) were present for the monthly point-prevalence surveys and were included in the final analysis.
Use of the three different antibiotic classes overall was similar during the two intervention periods, accounting for 42% of all antibiotics used in the ICUs during the cycling intervention and 43% of antibiotics used during mixing (compared with 39% during baseline). Microbiological screening results showed that the mean prevalence of antibiotic-resistant, gram-negative bacteria was 23% during the cycling interventions (168 out of 745 patients) and 22% during mixing (184 of 853 patients).
The authors of the study say the interventions achieved their intended goal of varying the use of the three antibiotics without changing the volume used over time. In addition, there were substantial differences in exposure to the three antibiotics in the two interventions. During cycling, antibiotic use for the non-preferred drugs fell by half to two-thirds, while use of the three antibiotics remained stable during mixing. But the differences in exposure didn't affect the overall prevalence of antibiotic resistance on the ICUs.
"In this cluster-randomised crossover study in eight ICUs, 9-month periods of antibiotic cycling and mixing did not change the unit-wide prevalence of antibiotic-resistant, gram-negative bacteria," the authors write. "Therefore, structured rotation of antibiotic prescription policies for possible gram-negative bacteria cannot be considered as a measure to reduce antibiotic resistance in ICUs."
Despite the results, lead study author Joppe van Duijn, MD, of the Julius Center for Health Sciences and Primary Care in the Netherlands said in an email that a lot of unanswered questions remain about antibiotic rotation. "There are a lot of rotation strategies possible, of which we tested just two," van Duijn said. "Antibiotic rotation will, therefore, need to be further studied to find out which specific intervention fits which specific ICU best."
Ultimately, van Duijn and his colleagues conclude, lowering the total volume of antibiotics used in ICUs remains the best strategy for reducing selection pressure and controlling the emergence of antibiotic-resistant bacteria. That can best be achieved, they say, by improved diagnostics that can determine which patients need antibiotics and which don't, and by biomarker-guided reductions in antibiotic duration.
Jan 24 Lancet Infect Dis study