Special Feature

Antibiotic Cycling in the ICU

By Jun Takezawa, MD


Once intensive care practices have been well standardized, patient safety becomes one of the biggest issues in the ICU in terms of management of health care quality (affecting the outcomes of patients admitted to the ICUs). The quality of health care in the ICU is related to 1) organizational characteristics (such as staffing, open vs closed systems, workforce); 2) standardization of treatment and care process; 3) competence of the ICU staff in providing treatment and care; and, 4) risk hedging capability against medical errors which include acquisition of nosocomial infections. The acquisition of nosocomial infections in the ICU is known to significantly affect patient outcomes such as ICU and hospital mortality and length of ICU and hospital stays.

Risk Factors for ICU-Acquired Infections

Environmental risk factors in the development of ICU-acquired infections include widespread use of broad-spectrum antibiotics, increased utilization of the ICU by more seriously-ill patients, increased use of invasive medical devices, lack of well-trained nurses and their increased workload, and the increased numbers of patients who stay in the ICU due to chronic and severe illness. The risk factors for ICU-acquired nosocomial infections, which are preventable, are insufficiency in hygiene practice, uncontrolled antibiotic usage, delays in administration of the appropriate antibiotic, and emergence of antibiotic-resistant organisms. Although hygiene practice is standardized by various guidelines published by the professional societies such as APIC and NCRQ and governmental health care agencies, the use of antibiotics is not well standardized, and the numbers of nosocomial infections caused by antibiotic-resistant organisms is still increasing.

Prevention of Emergence of Antibiotic-Resistant Organisms

Several attempts have been made to promote the judicious use of antibiotics in preventing the emergence of antibiotic-resistant organisms, in hopes of decreasing the number of ICU-acquired infections caused by the antibiotic-resistant organisms. One strategy is limiting or restricting the number of antibiotics used in the ICU. Although a large scaled randomized controlled trial in order to obtain a definitive conclusion on this issue would be difficult to do, it is rational to believe that a decrease in the total number of antibiotics used in a given hospital would result in improvement of the antibiotic susceptibility of the organisms which inhabit the hospital environment.

Antibiotic Cycling

Another strategy to reduce the emergence of antibiotic-resistant bacteria is the use of various antibiotic classes. This has been accomplished by alternating or cycling antibiotics used in the ICU during predetermined periods.

The background rationale for antibiotic cycling is that selective pressure by the antibiotic is reduced during the restricted period, and susceptibility to the antibiotics on the part of resistant organisms is improved. Additionally, the chance for emergence of resistant organisms may be decreased if susceptibility to the antibiotics remains stable. As far as susceptibility of the antibiotics being reversible, and the baseline susceptibility improved by restricting the use of a certain class of antibiotics, this strategy may be effective in reducing the average resistance rate of the antibiotics during cycling. Therefore, if the baseline susceptibility is not improved or does not return to the baseline level, this strategy is ineffective.

Reported Experiences with Antibiotic Cycling

The antibiotic cycling strategy was first introduced by Gerding et al1 where gentamicin and amikacin were cycled alternatively. This strategy was designed in response to increased resistance of Gram-negative bacteria to gentamicin. During the first 4-month period, in which gentamicin was predominantly used, resistance to gentamicin and amikacin was 12% and 3.8%, respectively. During the next 26 months, amikacin was predominantly used. Resistance to gentamicin decreased to 6.4%, but that to amikacin increased to 9.2%. Then gentamicin was re-introduced for the next 12 months. The resistance to gentamicin increased to 9.2%, while the resistance to amikacin was 3.9% which is the same resistance rate as was found in the pre-cycling period. Finally, amikacin was re-introduced for 12 months. The resistance to gentamicin and amikacin was 5.4% and 2.8%, respectively. These results suggested that antibiotic cycling reduced the resistance to the cycled antibiotic when the baseline resistance had already increased, and the resistance level remained stable as far as cycling was continued.

Young et al2 cycled the use of gentamicin and amikacin because of the increased resistance to gentamicin in 14 US hospitals. During the baseline period where gentamicin was primarily used and amikacin was restricted, resistance to gentamicin was 14% and that to amikacin was 2.4%. In the next 15 months, gentamicin was restricted and amikacin was used as a first line aminoglycoside. The resistance to gentamicin was significantly decreased to 9.2%, while resistance to amikacin remained at the same level of 2.2%. During the subsequent period when cycling and restriction were terminated, the resistance to gentamicin significantly increased to 15.3%, and that to amikacin was also significantly increased to 4%.

Because of the increased incidence of Gram-negative bacterial infections, especially for ventilator-associated pneumonia (VAP), Kollef et al3 predominantly used antibiotics from ceftazidime to ciprofloxacin for 6-month intervals in treating Gram-negative bacterial infections. They found that the incidence of VAP due to Gram-negative bacteria was significantly reduced when ciprofroxacin was used and ceftazidime was restricted.

Gruson et al4 controlled antibiotic use for the treatment of VAP by restricting both ceftazidime and ciprofloxacin for empiric and therapeutic uses. When the Gram-negative bacteria responsible for VAP were identified, cefepime, piperacillin-tazobactam, imipenem, and ticarcillin-clavulanic acid were sequentially changed for one-month intervals. They found that the number of antibiotic-resistant Gram-negative bacteria responsible for VAP was significantly reduced. The susceptibility of the gram-negative bacteria to this antibiotic regimen was also significantly improved. In addition, the prevalence of methacillin-sensitive strains of Staphylococcus aureus responsible for VAP was also increased from 40% to 60%.

Kollef et al5 subsequently rotated 3 different classes of antibiotics used for treating Gram-negative infections in a surgical and medical ICU consecutively for a 6-month period. They found that inadequate administration of antibiotics to nosocomial infections caused by Gram-negative bacteria was reduced and hospital mortality was significantly improved in patients whose APACHE scores were more than 15. However, they also found that the resistance of the bacteria to the antibiotics used during cycling was significantly increased compared to the resistance when they were restricted during the cycling periods.

Because of the increased incidence of colonization of vancomycin-resistant Enterococcus species (VRE) in hematological malignancy unit, Bradley et al6 rotated the use of antibiotics; ceftazidime and piperacillin-tazobactam. During the initial 4 moth period, where ceftazidime was used, the incidence of VRE colonization was 57%; during the subsequent 8-month period, in which ceftazidime was replaced with piperacillin-tazobactam, the incidence of VRE colonization was significantly decreased to 29%. When ceftazidime was introduced again, the incidence of VRE colonization returned to 36%.

Raymond et al7 rotated the empirical use of antibiotics every 3 months in the surgery-trauma ICU for one year. Ciprofloxacin+/-clindamycin, piperacillin/tazobactam, carbapenem, and cefepime+/-clindamycin were rotated, and the effects of antibiotic rotation on patient outcomes were evaluated in comparison with those during non-protocol driven antibiotic use for a period of one year. The incidence of Gram-negative bacterial infections and mortality associated with infection were significantly decreased by the antibiotic rotation. However, during the protocol driven period, formulary change from ceftazidime to cefepime occurred and both antibiotic surveillance and a new hygiene practice were started.

Toltzis et al8 conducted a clinical trial to evaluate whether antibiotic cycling decreased the incidence of colonization of multidrug-resistant Gram-negative bacteria in patients admitted to a neonatal intensive care unit. Study patients received empirical antibiotic therapy on a one-month rotation basis, employing gentamicin, piperacillin-tazobactam, and ceftazidime. The control patients received empirical antibiotics according to the preference of the physician. There was no significant difference in the amount of antibiotics used between the study and control patients. Also, no difference was found in the incidence of colonization of multidrug-resistant Gram-negative bacteria or in clinical outcome.

From these observations, antibiotic cycling was effective in reducing the incidence of colonization of antibiotic-resistant bacteria and ICU-acquired infections caused by antibiotic-resistant bacteria. The susceptibility of the cycling antibiotics was also improved during the withdrawal period. However, the rate of resistance to the initial antibiotics returned to the baseline level, when the same antibiotics were re-introduced.

Difficulties in Clinical Trials

There are several problems inherit to the antibiotic cycling strategy in conducting clinical trials:

  • The susceptibility of the antibiotics used in a certain ICU is different among the ICUs, and choice of antibiotics cycled as well as optimal duration of antibiotic cycling may be different among the ICUs. Thus, the positive result obtained from one ICU may not be applicable to other ICUs.
  • The baseline antibiotic susceptibility is always monitored to confirm that antibiotic cycling improves the average susceptibility of the cycled antibiotics during the whole cycling phases.
  • It always takes a long period of time to accomplish an antibiotic cycling trial, and during this period it is likely that many confounders such as new infection control practices and new devices could be introduced into clinical practice, which may affect the result of antibiotic cycling strategy.
  • It is highly possible that resistant bacteria can be introduced into ICUs as community-acquired or ward-acquired infections. This may affect the baseline susceptibility of the antibiotics used for cycling.
  • Although randomized, controlled trials can be organized to exclude confounders and biases associated with clinical trials, this type of clinical trial is expensive and labor-intensive, especially in view of the fact that, ideally, cycling antibiotics should be blinded.
  • The biggest problem is which outcome measure should be adopted, such as baseline resistance, incidence of ICU-acquired infections, ICU or hospital mortality, length of ICU or hospital stay, and cost.
  • To measure performance of infection control practice, a risk adjustment is required. However, risks for each outcome are different and it is not evaluated thoroughly.

Current Recommendation

Although antibiotic cycling seems to be effective in reducing the baseline resistance of the causative organisms which are responsible for ICU-acquired infections, this strategy has not been verified by clinical trials with any strong evidence. This strategy may be beneficial when the baseline resistance is already high. However, the baseline resistance should be carefully monitored, to confirm that it is stable or not increasing. The best practice for prevention of ICU-acquired infection is early and sufficient administration of optimal antibiotics to the infected patient, and more importantly, having both the incidence of ICU-acquired infections and the amount of antibiotics used in the ICU also low. In order to accomplish this goal, however, sound daily infection control practice is most important. Therefore, antibiotic-cycling is not recommended as a routine clinical practice for prevention of emergence of antibiotic-resistance bacteria and ICU-acquired infections caused by them.


1. Gerding DN, et al. Am J Med. 1985;79 (Suppl 1A):1-7.

2. Young EJ, et al. Am J Med Sci. 1985;290:223-227.

3. Kollef MH, et al. Am J Respir Crit Care Med. 1997;156:1040-1048.

4. Gruson D, et al. Am J Respir Crit Care Med. 2000;162:837-843.

5. Kollef MH, et al. Crit Care Med. 2000;28: 3456-3464.

6. Bradley SJ, et al. J Antimicro Chemother. 1999;43:261-266.

7. Raymond DP. Crit Care Med. 2001;29:1101-1108.

8. Toltzis P. Pediatrics. 2002;110:707-711.

Jun Takezawa, MD, Director of Emergency and Intensive Care Medicine Professor, Department of Emergency Medicine Nagoya University School of Medicine Nagoya, Japan, is Associate Editor for Critical Care Alert.