Antimicrobial resistance is a global threat to public health, increasing health care costs, treatment failures, and mortality.1 The critically ill patient admitted to the ICU has a greater vulnerability to infections and is particularly susceptible to be colonized/infected by multidrug-resistant bacteria (MDR-B).2,3 In Spain, infections related to invasive devices in critically ill patients have been collected for years through the National Nosocomial Surveillance Study (ENVIN-HELICS), allowing assessment of changing trends, etiology of individual infections controlled by the surveillance program, and susceptibility profile of the most frequent pathogens.4 ICU infections caused by MDR-B have similar clinical manifestations than those caused by susceptible bacteria, but treatment options are much more limited. Moreover, MDR-B infections are associated with inappropriate empirical antibiotic treatment that delays the onset of appropriate therapy, and therapeutic failure. As a consequence, mean ICU stay is prolonged, with an increase in costs, morbidity, and mortality.5,6

International organizations and scientific societies have developed strategies to reduce the emergence and spread of antimicrobial resistance, emphasizing on the need for the implementation of control measures.7–11 In Spain, the Spanish Society of Intensive Care Medicine and Coronary Units (SEMICYUC) and the Spanish Society of Intensive Nursing and Coronary Units (SEEIUC) have led two projects promoted by the Spanish Ministry of Health, Social Policy and Equality (MSSSI) aimed to reduce central venous catheter-related infections (“Bacteremia Zero” project)12 and ventilator-associated pneumonia (“Zero-VAP” project)13 in the ICU setting. The implementation of these projects has been associated with a 50% decrease of these infections and contributed to reduce antimicrobial consumption for the treatment of ICU-acquired infections. Using the structure already developed for these two projects, a new intervention project, called “Zero Resistance” was designed and implemented. The primary objective of the “Zero Resistance” project was to reduce acquisition of one or more MDR-B during the patients’ ICU stay, based on the hypothesis that a multimodal intervention or a set of specific measures applied simultaneous in daily practice can decrease the rate of colonization and/or infection by MDR-B in ICU patients. The project was approved by the Ethics Committee of ‘Parc de Salut Mar’ of Barcelona.

A total of 103 ICUs from 15 autonomous communities participated in the “Zero Resistance” project, with a total of 139,505 patients and 833,228 days of ICU stay. Most ICUs were polyvalent (88%) from public hospitals (99.6%) with pregraduate teaching programs (70.7%), and were located in large hospitals (>500 beds) (36%) or medium-size hospitals (200–500 beds) (47%). In 51 (49.5%) ICUs, surveillance cultures were performed on admission and weekly in all patients. In 34 (33%) ICUs, surveillance cultures were performed only on admission to the population with risk factors of being MDR-B carriers and weekly in all patients. In 17 (16.5%) ICUs, surveillance cultures were performed on admission and weekly exclusively to the population with risk factors, and in 1 ICU surveillance cultures were performed on admission in all patients and weekly to the target population.

Adherence to each recommendation of the “Zero Resistance” bundle is shown in Fig. 1. A total of 10,289 MDR bacteria were identified during the study period. Of these 10,289 MDR-B, 6020 (58.5%) were isolated in 5409 patients (3.9%) on ICU admission (colonized 2.8%, infected 1.2%) and 4269 (41.5%) were new MDR-B isolated in 3648 (2.6%) patients during their stay in the ICU (colonized 1.7%, infected 1.0%). In 287 patients, MDR-B identified during their ICU stay were different than bacteria identified on admission.

Figure 1.

Degree of adherence to the bundle recommendations of the “Zero Resistance” program.

(0.17MB).

Monthly trends of MDR-B rates of patients on ICU admission and during ICU stay are shown in Fig. 2. The rate of patients with MDR-B detected on admission increased significantly, with an incidence rate ratio (IRR) of 1.43 (95% CI 1.31–1.56) over the RZ project (p<0.001), with an increase of 32.2% between the initial and final monthly rates. On the contrary, the rate of patients with MDR during ICU stay decreased, although non-significantly, with an IRR of 0.93 (95% CI 0.83–1.03) throughout the study period (p=0.174), with a 24.9% decrease between initial and final monthly rates. According to the classification into colonization or infection (Figs. 3 and 4), MDR-B colonizations showed a highly significant increase on ICU admission (IRR 1.69, 95% CI 1.52–1.83; p<0.0001) and a non-significant increase during ICU stay (IRR 1.12, 95% CI 0.98–1.29; p=0.086), meanwhile MDR-B infections showed as a non-significant decrease on ICU admission (IRR 0.95, 95% CI 0.81–1.12; p=0.549) and a very significant decrease during ICU stay (IRR 0.67, 95% CI 0.57–0.80, p<0.0001) with a 46% decrease between initial and final monthly rates.

Figure 2.

Monthly smoothed trends of the rate of patients with one or more MDR-B identified on ICU admission or during their ICU stay.

(0.2MB).

Figure 3.

Monthly smoothed trends of the rates of patients with MDR-B identified on admission according to colonization or infection status.

(0.16MB).

Figure 4.

Monthly smoothed trends of the rates of patients with MDR-B identified during ICU stay according to colonization or infection status.

(0.16MB).

Fig. 5, shows the distribution of MDR bacteria depending on the moment of their identification. Enterobacteriacea producing extended-spectrum beta-lactamases (ESBLs) and methicillin-resistant Staphylococcus aureus (MRSA) were more frequently isolated at the time of ICU admission, whereas imipenem-resistant Acinetobacter baumannii, Pseudomonas aeruginosa and carbapenemase-producing Gram negative bacilli were more frequently isolated during ICU stay. The distribution of individual MDR-B in colonized and infected patients on ICU admission and during ICU stay is shown in Table 3. There were significant differences in the distribution in MDR-B specific infections depending on the moment of their identification, with more presence of infections by imipenem-resistant Acinetobacter baumannii, Enterobacteriaceae producing ESBLs and MRSA in MDR-B isolated during the ICU stay than at ICU admission.

Figure 5.

Distribution of different MDR-B according to isolation on ICU admission or during the patients’ ICU stay.

(0.15MB).

The rate of days of isolation for preventing MDR-B spread increased significantly during implementation of the “Zero Resistance” program with an IRR of 1.42 (95% CI 1.40–1.44) (p<0.001) and a mean of 257.58 per 1000 days of ICU stay.

The implementation of the “Zero Resistance” program in Spanish ICUs has been associated with a non-significant decrease in the rate of patients with MDR-B acquired during ICU stay taking together colonizations and infections, but it was associated with a very significant decrease in the rate of patients with infection produced by MDR-B acquired during ICU stay (IRR=0.67). Moreover, the “Zero Resistance” program has demonstrated that the number of MDR bacteria identified at the time of admission was significantly higher than those acquired during the patients’ ICU stay.

There are no previous reports of the implementation of a bundle for preventing MDR-B infection in the ICU setting as those recommended in the present “Zero Resistance” project. To date, control interventions alone o combined have been proposed, such as to reduce or limit excessive antibiotic prescribing and/or antibiotic cycling,17 to preform surveillance cultures for early detection of MDR-B,18–21 to use barrier precautions in patients at risk for MDR-B carrier status,22,23 and search and destroy strategies for eradication of reservoirs.24,25 Antimicrobial stewardship programs have shown to reduce antimicrobial consumption and costs26,27 as well as length of hospital stay and mortality.28 However, there is no evidence of the effect of antimicrobial stewardship programs in the prevention of MDR-B infections in ICU patients.29 The use of strategies to identify MDR-B carries has been mostly focused on MRSA.18–22,30 Results of a systematic review showed that active surveillance cultures on ICU admission and weekly thereafter reduced the incidence of MRSA infection, but the overall quality of the evidence was poor.18 In Spain, the Spanish Society of Infectious Diseases and Clinical Microbiology (SEIMC) and the Spanish Society of Preventive Medicine Public Health and Hygiene (SEMPSPH) recommended systematic screening cultures for MRSA during the first hours of ICU admission.31 On the other hand, a few studies of pre-emptive contact precautions to reduce health care (HA) — associated MRSA infections have been shown to be effective in reducing the incidence of HA-MRSA transmission.32,33

In recent years, there has been also an increase of ESBL-producing and carbapenemase-producing Gram negative bacteria, multiresistant P. aeruginosa, or imipenem-resistant A. baumannii among hospitalized patients and critically ill patients admitted to the ICU.3,34 The present findings confirm that, in many cases, patients are carriers of these microorganisms on ICU admission, so that different strategies based on the presence of risk factors have been developed to support preventive isolation measures.14,35,36 However, criteria to select these patients are not sufficiently sensitive or specific, so that many patients without MDR-B are isolated and, on the other hand, MDR-B are identified in patients in which criteria are not met. Also, implementation of antibiotic optimization policies based on surveillance cultures on ICU admission and preventive isolation measures may be ineffective in the presence of MDR-B reservoirs. The need to search systematically reservoirs and to develop cleaning protocols have been extensively recommended, particularly to identify potential sources and eradication of reservoirs as key interventions to achieve termination of MDR-B outbreaks in ICUs.37

MRD bacteria were most frequently identified in surveillance samples taken on ICU admission, especially noticeable throughout the implementation period, suggesting that the main reservoir of MDR bacteria is outside the ICU, which contradicts the accepted concept of considering ICUs as the epicenter of MDR-B development.5 The active search of MDR bacteria on ICU admission was associated with a significant increase in MDR-B isolates classified as colonization with an estimated incidence rate at the end of the study 1.69 times the baseline one. It is possible that for years, many MDR-B isolated in the ICU and related to the use of antibiotics or cross-transmission would have been already present in patients on ICU admission. At the same time, whereas a significant reduction in MDR-B infections on ICU admission was not observed, ICU-acquired MDR-B infections decreased markedly throughout implantation of the “Zero Resistance” program with an IRR of 0.67. Early identification of MDR-B involves the use of products containing 2%–4% chlorhexidine in daily patient hygiene, which reduces MDR-B spread and the risk of epidemic outbreaks.38–40

A very remarkable result of the study was the detection of important differences in the distribution of MDR-B on admission and during ICU stay. ESLP-producing Enterobacteriaceae, and MRAS were especially imported to the ICUs (colonization pressure), while MDR P. aeruginosa, carbapenemase-producing Gram negative bacilli and imipenem-resistant A. baumannii, appeared preferentially during ICU stay, which suggest the presence of reservoirs for these microorganisms in the Spanish ICUs.

Limitations of the study include possible methodological differences among ICUs in surveillance sampling, bacterial culture, identification of pathogens, and susceptibility testing. Bias related to interobserver variability in the collection and registration of clinical variables cannot be excluded. Also, recommendations could have been differently applied among the participating ICUs, particularly regarding sampling strategy on admission and during ICU stay.

As a summary of the study, it can be concluded that implementation of ZR project-recommendations was associated with a non-significant decrease in the rate of patients with MDR-B acquired during ICU stay taking together colonizations and infections, but it was associated with a very significant decrease in the rate of patients with infection produced by MDR-B acquired during ICU stay. At the same time, a significant increase of patients with MDR-B on ICU admission was observed, which allowed the implementation of isolation measures to prevent dissemination. It is possible to reduce infections caused by MDR-B in the ICU by means of a multiple intervention focused on the use of antibiotics, organized search of MDR-B on admission and during ICU stay, and elimination of reservoirs of MDR-B in the ICUs.

F. Álvarez-Lerma: project design and coordination, patient recruitment, statistical analysis, methodological review, interpretation of results, and writing of the manuscript. M. Catalán-González: project design and coordination, patient recruitment, methodological review, interpretation of results and writing of the manuscript. J. Álvarez, M. Palomar-Martínez, I. Fernández-Moreno, J. Garnacho-Montero, F. Barcenilla-Gaite, R. García, and J. Aranaz-Andrés: project design and coordination, patient recruitment, methodological review, interpretation of results. M. Sánchez-García: project design and coordination, patient recruitment, interpretation of results and writing of the manuscript. F.J. Lozano-García: project coordination, and interpretation of results. M. Martínez-Alonso: statistical analysis, methodological review, interpretation of results and writing of the manuscript. All authors have seen and approved the final draft.

Supported by the Spanish Ministry of Health, Social Policy and Equality through a contract (number 2014/407PN010).