Infections and antimicrobial resistance in the intensive care unit: Epidemiology and prevention

INTRODUCTION — Although intensive care units (ICUs) account for fewer than 10 percent of total beds in most hospitals, more than 20 percent of all nosocomial infections are acquired in ICUs [1]. ICU-acquired infections account for substantial morbidity, mortality, and expense. Infections and sepsis are the leading cause of death in noncardiac ICUs and account for 40 percent of all ICU expenditures [2].

The epidemiology of nosocomial ICU infections and antimicrobial resistance in ICUs will be discussed here. The most important nosocomial infections in the ICU, namely catheter-related bloodstream infections, ventilator-associated pneumonia, and catheter-associated urinary tract infections, will be discussed briefly here and in more detail separately:

●(See "Catheter-associated urinary tract infection in adults" and "Complications of urinary bladder catheters and preventive strategies".)

●(See "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults" and "Treatment of hospital-acquired and ventilator-associated pneumonia in adults" and "Risk factors and prevention of hospital-acquired and ventilator-associated pneumonia in adults".)

(See "Intravascular catheter-related infection: Epidemiology, pathogenesis, and microbiology" and "Intravascular non-hemodialysis catheter-related infection: Treatment".)

Issues related to surgical site infection are discussed separately. (See "Antimicrobial prophylaxis for prevention of surgical site infection in adults" and "Overview of control measures for prevention of surgical site infection in adults".)

Infection prevention and ICU care of patients with COVID-19 are also presented elsewhere. (See "COVID-19: Infection prevention for persons with SARS-CoV-2 infection" and "COVID-19: Management of the intubated adult".)

EPIDEMIOLOGY

Prevalence of infections in the ICU — Despite the advances in modern medicine and intensive care, the incidence of sepsis in intensive care units (ICUs) continues to rise. In an international study of 1265 ICUs, 60 percent of ICU patients at the time of survey were considered infected, with infection being a strong independent predictor for mortality (odds ratio [OR] 1.51, p<0.001) [2]. The risks of infection in general and with a resistant pathogen in particular increased with the length of patient stay in the ICU. Several factors contribute to the high incidence of these infections in the ICU and the associated poor patient outcomes:

●Compared with patients in the general hospital population, patients in ICUs have more chronic comorbid illnesses and more severe acute physiologic derangements and thus are relatively immunosuppressed [3].

●The high frequency of indwelling catheters among ICU patients provides a portal of entry of organisms into vital body organs and sites. The use and maintenance of these catheters necessitate frequent contact with health care personnel, which predispose patients to colonization and infection with nosocomial pathogens. In addition, equipment associated with the proper maintenance of these devices might serve as reservoirs and vectors for pathogens and be related to horizontal patient-to-patient transmission of pathogens [4].

●Multidrug-resistant pathogens, such as methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), Acinetobacter baumannii, Enterobacteriaceae that produce extended-spectrum beta-lactamases and/or carbapenemases (eg, ESBLs and CREs, respectively), and carbapenem-resistant Pseudomonas aeruginosa, are all being isolated with increasing frequency in ICUs [5,6]. Infections caused by these resistant pathogens are difficult to treat and are associated with increased morbidity, mortality, and costs [7,8]. (See 'Prevalence of drug-resistant organisms' below.)

●Compared with patients in the general hospital population, patients in ICUs are subjected to increased selective pressure and increased colonization pressure [3,9].

Although most studies of ICU-associated infections come from industrialized countries, the rates of infection may even be higher in developing countries, as illustrated by a multicenter prospective cohort surveillance study of 46 hospitals in Central and South America, India, Morocco, and Turkey [10]. An overall rate of 14.7 percent (or 22.5 infections per 1000 ICU days) was observed. The following rates were found for specific devices:

●Ventilator associated pneumonia (VAP); 24.1 cases per 1000 ventilator days (range 10.0 to 52.7 cases)

●Catheter-related bloodstream infection (CRBSI); 12.5 cases per 1000 catheter days (range 7.8 to 18.5 cases)

●Catheter-associated urinary tract infections (CAUTI); 8.9 cases per 1000 catheter days (1.7 to 12.8 cases)

A subsequent study by the same international group had reported results from 98 ICUs from Latin America, Asia, Africa, and Europe [11]. Despite the fact that device utilization was remarkably similar to that reported from ICUs in the United States, rates of device-associated nosocomial infection were markedly higher in the ICUs from the developing world.

Common infectious syndromes in the ICU — The most common and clinically important infections acquired in the ICU are those associated with the supportive devices that patients in the ICU often require. These include intravascular catheter-related bloodstream infection, ventilator-associated pneumonia, and catheter-associated urinary tract infection. Issues related to the management of severe COVID-19 infections requiring ICU care are discussed elsewhere. (See "COVID-19: Management in hospitalized adults" and "COVID-19: Management of the intubated adult".)

Catheter associated urinary tract infection — Urinary tract infection (UTI) is the most common nosocomial infection, accounting for more than 40 percent of all nosocomial infections [12]. While most catheter-associated UTIs do not cause severe morbidity and mortality or significantly increase hospital costs, the cumulative impact of these frequent infections is large [13]. In the United States, CAUTIs are responsible for 900,000 additional hospital days per year and contribute to >7000 deaths [14,15]. CAUTIs are the second most common cause of nosocomial bloodstream infection (ie, urosepsis), which have an attributable mortality of approximately 15 to 25 percent [16-20]. The Centers for Medicare and Medicaid Services (CMS) does not reimburse hospitals for CAUTIs, which further increases the cost burden of these infections on hospitals in the United States.

In addition to actual infection, asymptomatic bacteriuria often leads to significant laboratory testing and inappropriate antimicrobial utilization in the absence of an established infection [19,21]. Inappropriate treatment of asymptomatic bacteriuria has been associated with adverse clinical outcomes [22]. The urinary tract in catheterized patients also serves as a reservoir for multidrug-resistant bacteria, which can cause either infection or asymptomatic bacteriuria [19]. Issues related to CAUTIs are discussed separately. During the work-up of a patient for suspected infection, if a non-urinary source of infection is suspected, routine urine culture should be avoided. (See "Catheter-associated urinary tract infection in adults" and "Placement and management of urinary bladder catheters in adults".)

Ventilator associated pneumonia — Ventilator-associated pneumonia is infection of lung tissue that develops 48 hours or more after intubation in mechanically ventilated patients. Nosocomial pneumonia is the second most common hospital-acquired infection and occurs frequently in the setting of endotracheal intubation and mechanical ventilation [23]. These issues are discussed in detail separately. (See "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults" and "Risk factors and prevention of hospital-acquired and ventilator-associated pneumonia in adults" and "Treatment of hospital-acquired and ventilator-associated pneumonia in adults".)

Intravascular catheter-related bloodstream infection — Arterial and central venous catheters are frequently used in critical care patients because of the need for hemodynamic monitoring and intravenous therapeutics. Bloodstream infections involving these catheters are common in ICUs and are associated with significant morbidity and mortality [24]. In addition, in the United States, the cost burden of these infections on health care facilities was exacerbated after the Centers for Medicare and Medicaid Services stopped reimbursing hospitals for catheter-related bloodstream infections in October 2008. These infections are discussed in detail elsewhere. (See "Intravascular catheter-related infection: Epidemiology, pathogenesis, and microbiology" and "Intravascular non-hemodialysis catheter-related infection: Clinical manifestations and diagnosis" and "Intravascular non-hemodialysis catheter-related infection: Treatment".)

Prevalence of drug-resistant organisms — There has been a rapid rise in the rate of resistance among bacterial pathogens recovered in intensive care units. A comparison of the reports from the National Healthcare Safety Network System (NHSN) at the United States Centers for Disease Control and Prevention (CDC) from 1999 and 2006 to 2007 demonstrates increasing prevalence of multidrug-resistant pathogens in ICUs in the United States [5,6]:

●VRE (from 24.7 to 33.3 percent of enterococci isolates)

●MRSA (from 53.5 to 56.2 percent of S. aureus isolates)

●P. aeruginosa resistant to imipenem or fluoroquinolones (from 16.4 to 25.3 and from 23.0 to 30.7 percent of P. aeruginosa isolates, respectively)

●A. baumannii resistant to carbapenems (from 11 to 30 percent of A. baumannii isolates)

●Enterobacteriaceae resistant to third-generation cephalosporins, mainly ESBL producers (from 10.4 to 25 percent of Klebsiella pneumoniae and 3.9 to 9 percent of Escherichia coli isolates) (see "Extended-spectrum beta-lactamases", section on 'Epidemiology')

●Enterobacteriaceae resistant to carbapenems (CREs) (from 0 to 8 percent of K. pneumoniae and from 0 to 3 percent of E. coli)

The emergence of broad-spectrum resistance among gram-negatives is particularly worrisome since therapeutic options are scarce, and sometimes no effective antimicrobial agent is available at all [25]. (See "Extended-spectrum beta-lactamases", section on 'Treatment options' and "Overview of carbapenemase-producing gram-negative bacilli", section on 'Treatment' and "Principles of antimicrobial therapy of Pseudomonas aeruginosa infections", section on 'Management of multidrug-resistant organisms'.)

Two additional common and significant pathogens in ICU infections are Clostridioides difficile and Candida spp. These are not "traditional" multidrug-resistant organisms (MDRO), but risk factors for infections due to these pathogens are similar to those associated with MDRO infections, and thus the affected populations are similar [26,27]. These pathogens are discussed in detail elsewhere. (See "Clostridioides difficile infection in adults: Clinical manifestations and diagnosis" and "Clostridioides difficile infection in adults: Treatment and prevention" and "Overview of Candida infections".)

Risk factors for resistant infections — Certain characteristics increase the risk of infections with multidrug-resistant pathogens in ICUs by contributing to increased selective pressure (leading to the emergence of multidrug-resistant organisms) and/or increased colonization pressure (leading to ineffective containment of these organisms) [9,28,29]. Specifically, risk factors for resistant infections reported from ICUs include the following [2,30-35]:

●Older age

●Lack of functional independence and/or decreased cognition

●Presence of underlying comorbid conditions (eg, diabetes, renal failure, malignancies, immunosuppression) and higher severity of acute illness indices

●Long duration of hospitalization prior to the ICU admission, including interinstitutional transferring (particularly from nursing homes)

●Frequent encounters with health care environments (eg, hemodialysis units, ambulatory daycare clinics)

●Frequent contact with health care personnel concurrently caring for multiple patients, whose hands can serve as vehicles for transfer of pathogens between patients. Shared equipment and contaminated environments can also serve as reservoirs and/or vectors that contribute to acquisition of infections in the ICU.

●Presence of indwelling devices such as central venous catheters, urinary catheters, and endotracheal tubes, which bypass natural host defense mechanisms and serve as portals of entry for pathogens

●Recent surgery or other invasive procedures

●Receipt of antimicrobial therapy prior to the ICU admission, which creates selective pressure promoting the emergence of multidrug-resistant bacteria

The association between prior receipt of antibiotics and infection with drug-resistant organisms has been demonstrated in several studies and by various methodologies. In case-control studies, exposure to antibiotics has consistently been associated with the emergence of resistance to that same or a different class of antimicrobial agent [36]. As an example, receipt of fluoroquinolones has been linked to the emergence of piperacillin-resistant P. aeruginosa [37]. In a study of patients with ventilator-associated pneumonia, those infected with piperacillin-resistant strains of P. aeruginosa were more likely to have received fluoroquinolones prior to their pneumonia (OR 4.6, 95% CI 1.7-12.7). In a separate study, antibiotic exposure was the strongest single predictor for infection with extensively drug-resistant gram-negative pathogens [36]. The association between certain antibiotic use and emergence of resistance has also been supported by studies that used longitudinal time-series analyses to determine rates of resistance when various antibiotic agents were more commonly used at a particular institution [38] and studies that demonstrate reductions in multidrug-resistant pathogens with the implementation of antimicrobial stewardship programs and various strategies to minimized unnecessary antimicrobial use [39].

Outcome of multidrug-resistant infections — Infections caused by multidrug-resistant pathogens are associated with increased mortality, length of hospital stay, and hospital costs [40-46]. Patients with infections due to multidrug-resistant organisms usually are chronically or acutely ill and at risk of dying from underlying serious and complex medical illnesses. However, a number of factors related to the difficulties of choosing antibiotics for multidrug-resistant bacteria independently predispose to poor outcomes. These include the following:

●Multidrug-resistant pathogens are more frequently resistant to empiric antimicrobial regimens than are susceptible organisms. Thus, there are often delays in initiation of appropriate, effective antimicrobial therapy in the treatment of multidrug-resistant organisms [47]. These delays are independent predictors of mortality in severe sepsis and thus contribute to the increased mortality rates associated with resistant infections [42,48-53]. As an example, in a study of patients with septic shock, each hour of delayed appropriate therapy in the first six hours of infection was associated with an average decrease in survival of 7.6 percent [54]. (See "Evaluation and management of suspected sepsis and septic shock in adults", section on 'Choosing a regimen'.)

●Antimicrobial resistance often precludes the use of optimal "first-line" antimicrobial agents and necessitates the use of "second-line" agents with inferior bactericidal activity and unfavorable pharmacokinetic and/or pharmacodynamic properties [50]. When "second line" agents are required to treat a resistant organism, adverse patient outcomes sometimes result [44,45,55-59]. As an example, vancomycin is commonly used to treat MRSA since anti-staphylococcal penicillins (eg, nafcillin) and first-generation cephalosporins (eg, cefazolin) are not active against the organism. However, vancomycin does not possess strong bactericidal activity and is associated with an increased risk for renal insufficiency compared with beta-lactams. In several clinical studies, vancomycin was inferior to beta-lactam agents in treating methicillin-susceptible S. aureus infections [60]. (See "Methicillin-resistant Staphylococcus aureus (MRSA) in adults: Treatment of bacteremia".)

Another factor that may contribute to the poor outcomes among patients with infections due to certain multidrug-resistant pathogens is the virulence properties of the organism, but this issue is subject to continuous debate [61].

PREVENTION — Strategies to prevent the emergence and spread of multidrug-resistant bacteria in intensive care units (ICUs) can be divided into two major categories: strategies that attempt to improve the efficacy and utilization of antimicrobial therapy (reducing selective pressure) and infection control measures (reducing colonization pressure) [50,62-64]. It is most efficacious to combine the two approaches [62].

Antibiotic utilization controls — Incorporating antibiotic stewardship programs (often involving clinicians, infectious diseases experts, and pharmacists) into specific hospital settings such as the ICU can comprehensively address the goal of reducing infections of resistant bacterial strains. The programs promote the effective and safe use of antimicrobial agents, evaluate and guide formulary decisions, and implement educational programs to improve antimicrobial utilization [64-68]. The specifics of antibiotic stewardship programs are discussed in detail elsewhere. (See "Antimicrobial stewardship in hospital settings".)

Antimicrobial stewardship programs in the ICU setting have been impactful and associated with decreases in drug-resistant bacteria in some study settings. As an example, in a study of two ICUs in the United States that implemented a comprehensive antimicrobial stewardship program, the proportion of hospital-acquired infections (HAI) caused by certain multidrug-resistant gram-negative bacilli, including P. aeruginosa, A. baumannii, and extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae decreased from 37.4 percent in 2001 to 8.5 percent in 2008 [39]. The rate of HAIs per 1000 patient-days that were caused by these multidrug-resistant organisms declined by 0.78 per year. Similarly, in a study of an ICU in Melbourne, Australia, which implemented an antimicrobial stewardship program, 2838 gram-negative bacilli were isolated from clinical cultures over seven years, and, over this time, there were significant increases in susceptibility of P. aeruginosa to imipenem (18.3 percent/year, p = 0.009) and gentamicin (11.6 percent/year, p = 0.02) compared with trends recorded prior to the stewardship program [69]. Improvements in the rates of gentamicin and ciprofloxacin susceptibility were also noted among Enterobacter spp.

There may be concerns that antibiotic stewardship may result in a delay in the initiation of appropriate antimicrobial therapy, which has been associated with poor clinical outcomes with infections caused by multi- and extensively drug-resistant infections [47,48]. However, a meta-analysis of five studies found that implementation of antimicrobial stewardship programs in ICUs was not associated with increased mortality (pooled relative risk 1.03, 95% CI 0.93-1.14), thus providing some reassurance that there is no clear evidence of unintended deleterious effects of stewardship programs on mortality in the ICU setting [70].

There is no role for rotating antibiotic prescription practices by changing empiric regimens in an attempt to curb emergence of resistance [71,72].

Infection control measures — Strategies to prevent the emergence of multidrug-resistant organisms that do not involve changes in antimicrobial utilization (which impacts selective pressure) primarily involve infection control measures (which impact colonization pressure and patient-to-patient transmission). Careful attention to these activities has been used to contain outbreaks of resistant organisms [73-76]. Adherence to hand hygiene, daily bathing with chlorhexidine, and implementation of device-specific strategies to decrease infection should be performed on a routine basis for all patients in an ICU [63]. For targeted reduction of methicillin-resistant S. aureus (MRSA), intranasal mupirocin can be used [77]. Contact precautions are warranted for patients infected or colonized with resistant organisms, patients with wound drainage that cannot be contained by dressings, and for patients with diarrhea. Surveillance for drug-resistant organisms is also important for identification and control of epidemic and endemic rates of resistance. Enhancing the regulation, the monitoring, and the processes for environmental cleaning is an additional established measure to contain the spread of multidrug-resistant organisms in the ICU [78]. New additive innovative technologies are currently available [79]. Many of these preventive measures, as well as strict adherence to standard precautions and regulations, are also effective in preventing severe acute respiratory system coronavirus 2 (SARS-CoV-2) transmission in ICUs [80]. (See "COVID-19: Infection prevention for persons with SARS-CoV-2 infection".)

Hand hygiene — There is no substitute for good hand hygiene compliance. Alcohol-based hand hygiene is more effective than traditional soap and water in cleansing hands of bacteria; in addition, no sink or towels are necessary, and alcohol foam is no more abrasive to hands than standard antiseptic soap and water [81,82]. Alcohol gel/foam is not appropriate for hands that are visibly soiled or for health care personnel caring for patients with C. difficile infection (or other spores-forming organisms), since the foam does not inactivate C. difficile toxins and does not kill the spores themselves. (See "Infection prevention: Precautions for preventing transmission of infection", section on 'Hand hygiene'.)

Contact precautions, cohorting, and dedicated staff — Wearing gown and gloves when entering a patient room and removing them prior to or shortly after exiting (but still adjacent to patient's close environment) may decrease transmission of multidrug-resistant bacteria, including MRSA, vancomycin-resistant enterococci (VRE), and carbapenem-resistant and ESBL-producing gram-negative organisms. These precautions should be routinely implemented when caring for ICU patients who have a history of or are found to have infection or colonization with resistant organisms. Evidence supporting this practice is discussed elsewhere. (See "Vancomycin-resistant enterococci: Epidemiology, prevention, and control", section on 'Contact precautions' and "Extended-spectrum beta-lactamases", section on 'Infection control and antibiotic stewardship' and "Methicillin-resistant Staphylococcus aureus (MRSA) in adults: Prevention and control".)

The issue of whether to use universal contact precautions for every patient in the ICU, regardless of colonization history, is a matter of ongoing debate. Although this may be a reasonable practice in outbreak settings or in institutions that have a high rate of colonization or infection with drug-resistant bacteria, routine use of universal contact precautions is not yet supported by strong scientific evidence. Some observational studies have suggested a decrease in transmission rates of drug-resistant organisms with universal contact precautions [83-85]. However, a large cluster-randomized trial failed to demonstrate a statistically significant benefit of universal use of contact precaution measures [86]. In this trial, 20 medical and surgical ICUs were randomly assigned to practice universal gown and glove use for all patients (intervention) or standard care with gown and glove use for only those patients known to be infected or colonized with antibiotic-resistant bacteria (control). Overall, 26,180 patients were followed over the three-month baseline period prior to implementation of the intervention and the 10-month study period following it. Compliance with gown and glove use was high, ranging from 80 to 86 percent. The intervention did not reduce the primary combined endpoint of MRSA or VRE acquisitions compared with baseline rates to a greater extent than the control (reduction from 21.4 to 16.9 versus 19.0 to 16.3 acquisitions per 1000 patient days in the intervention and control groups, respectively). The intervention did reduce MRSA acquisitions by 2.98 acquisitions per 1000 patient days (95% CI 0.38-5.58) more than the control; however, the baseline rate of MRSA acquisition was higher in the intervention group, so acquisition rates at the end of the study period were comparable between the two groups. The effect on transmission of other drug-resistant organisms was not evaluated.

In the trial, universal gown and glove use led to one fewer health care personnel visits per patient on average, but there was no excess of adverse patient events with this practice. This is in contrast with earlier observational studies that had reported adverse effects associated with implementation of universal contact precautions (eg, increased rates of falls, pressure ulcers, dissatisfaction of patients with their treatment, and less documented care safety) [87,88]. Long-term consequences of contact isolation precautions are yet unknown.

In another multicenter trial, contact isolation practices were observed among health care workers; as the proportion of patients in contact isolation increased, compliance with contact isolation precautions decreased [89]. A threshold for staff compliance (approximately 40 percent of patients on contact precautions) was noted.

Gowns and gloves should always be worn while caring for patients with wound drainage that cannot be contained by dressings. In addition, gown and gloves should always be removed prior to or immediately after leaving a patient's room/unit.

An additional potential control measure is geographically cohorting carriers of the same multidrug-resistant organism (MDRO) and assigning dedicated nursing staff to such patients. In one outbreak of carbapenem-resistant Enterobacteriaceae, the outbreak was contained only after implementation of dedicated cohorting [90]. However, a single unit for care of patients with multiple different types of MDROs should be avoided, since genes conferring resistance can cross between species [91].

Decolonization/patient bathing — We suggest daily chlorhexidine bathing for all ICU patients. Bathing patients daily with chlorhexidine gluconate (CHG), an antiseptic agent with broad-spectrum activity against many organisms, is an effective method of decreasing both hospital-acquired infections (ie, bloodstream infections, urinary tract infections, surgical-site infections, and ventilator-associated pneumonia) and colonization with drug-resistant organisms among patients in the ICU, as demonstrated in many studies [77,92-106]. Despite limitations of some of these studies, our recommendation is based on the apparent benefits, the low rate of associated adverse effects, and the relative ease of implementation. If MRSA is a specific concern, then intranasal mupirocin can be added. The use of targeted decolonization for patients with MRSA is discussed in detail elsewhere. (See "Methicillin-resistant Staphylococcus aureus (MRSA) in adults: Prevention and control", section on 'Targeted decolonization'.)

Chlorhexidine-impregnated cloths or chlorhexidine soaked washcloths should be firmly massaged over all patient body surfaces and skin folds below the jaw line except for the face. Most larger trials that evaluated chlorhexidine bathing used this approach with impregnated cloths [77,96]. It is uncertain whether bathing with a washcloth soaked with liquid chlorhexidine would result in similar effects, but this may be more accessible and less expensive than premade cloths. If a soaked washcloth is used, large amounts of water should be avoided (in order to retain an effective CHG concentration [ie, approximately 4 percent]), and care should be taken to avoid getting catheter dressing wet, which may be associated with an increased rate of catheter exit-site infections [107]. Reusable basins have been demonstrated to be a reservoir for multidrug-resistant organisms, but their role in the spread of such pathogens in the clinical settings are yet unclear [108].

Although slightly mixed, studies on chlorhexidine bathing generally support its efficacy. In a meta-analysis of two controlled trials and 10 observational studies of ICU patients, daily chlorhexidine bathing was associated with a decrease in health care-associated bloodstream infections compared with soap and water or no bath (odds ratio [OR] 0.44, 95% CI 0.33-0.59) [95]. Similarly, in a subsequent trial that included over 7000 patients in ICUs and bone marrow transplant units, there was a 28 percent reduction in the rate of hospital-acquired bloodstream infections (4.8 versus 6.6 cases per 1000 patient-days, respectively) [96]. The reduction in bloodstream infections with chlorhexidine was greatest for coagulase-negative staphylococci and fungal infections. Some studies suggest a trend towards reduced gram-negative infections [109]; however, more adequately powered, controlled studies are needed to better elucidate the effect on infection from specific organisms.

Chlorhexidine bathing has also been studied in critically ill pediatric patients. In a large trial including more than 6000 pediatric ICU patients older than two months, there was a nonsignificant reduction in the incidence of bacteremia with daily chlorhexidine bathing compared with standard bathing practices in the intention-to-treat analysis (3.5 versus 4.9 per 1000 days) [102]. In the per-protocol population, the incidence of bacteremia was lower among patients who received chlorhexidine bathing (3.3 versus 4.9 events per 1000 days, adjusted relative risk 0.64, 95% CI 0.42-0.98).

Although one cluster randomized, crossover study including 9340 adults in ICUs noted that daily chlorhexidine bathing did not demonstrate a reduced incidence of health care-associated infections, the study was underpowered to detect such differences due to the rarity of events, which limits the generalizability of the results [110,111]. A Cochrane systematic review that included this trial and seven others, comprising greater than 24,000 patients, noted reductions in rate of nosocomial infection (rate difference 1.70 fewer infections per 1000 patient-days, 95% CI 0.12-3.29) and mortality (OR 0.87, 95% CI 0.76-0.99) with chlorhexidine bathing, but deemed the evidence to be of very low quality [112].

Universal decolonization with chlorhexidine bathing and twice-daily mupirocin is a strategy that can be used to decrease MRSA bloodstream infections in the intensive care unit. In a large, multicenter trial that involved over 74,000 patients, universal decolonization with chlorhexidine and twice daily intranasal mupirocin reduced both MRSA-positive clinical cultures and bloodstream infections due to any pathogen (hazard ratios [HRs] 0.63 and 0.56 compared with baseline rates) to a greater extent than screening and isolation (HRs 0.92 and 0.99) or targeted decolonization of carriers (HRs 0.75 and 0.78) [77]. The number of patients requiring decolonization to prevent one MRSA infection or one bloodstream infection was 181 or 54, respectively. However, since both chlorhexidine and mupirocin were used for decolonization, the clinical effect in this study cannot be reliably attributed to chlorhexidine alone. Although, the Agency for Healthcare Research and Quality toolkit recommends universal decolonization with chlorhexidine and mupirocin [113], it is not clear to us that it is more effective than chlorhexidine alone, and so we do not routinely recommend it (except in cases of a targeted intervention to reduce S. aureus acquisitions and/or infections).

Reported adverse effects of chlorhexidine bathing are rare and predominantly mild skin reactions [95,102]. It is also relatively inexpensive.

However, emergence of resistance to chlorhexidine is an important consideration [97,114]. In a study including two 15-bed intensive care units over a four-year period, introduction of a chlorhexidine-based surface antiseptic protocol was associated with a 70 percent reduction in MRSA transmission, although transmission of strains carrying the plasmid-born qacA/B gene (which codes for multidrug efflux pumps and can lead to chlorhexidine resistance) was not reduced [97]. Resistance to triclosan, an ingredient in some antimicrobial soaps, has also emerged among dermal, intestinal, and environmental microorganisms, including S. aureus [115,116].

Digestive and oropharyngeal decontamination — Decontamination of the digestive and oropharyngeal tracts has been proposed as a method to reduce infection in critically ill patients by reducing microorganism colonization at these sites.

Modest mortality benefits have been demonstrated among ICU patients treated with selective oropharyngeal decontamination (SOD) and selective digestive decontamination (SDD) in the Netherlands, a region with low baseline antimicrobial resistance. However, these interventions have not found widespread favor outside the Netherlands, since thus far, no benefit for their use has been observed in ICUs with moderate to high levels of antibiotic resistance [117-119]. In addition, there has been uncertainty regarding the long-term effects of SOD and SDD on emergence of antimicrobial resistance [120-123].

Several studies have demonstrated that SOD and SDD each reduce mortality and rates of bacteremia among patients in ICUs with low levels of antibiotic resistance; the overall benefit is modest. Decontamination methods include oropharyngeal decontamination with antiseptics (eg, chlorhexidine), SOD with nonabsorbable antibiotics applied in the oropharynx, and SDD with nonabsorbable antibiotics applied to the oropharynx and administered orally, with or without intravenous antibiotics. Studies include:

●In a meta-analysis of randomized trials evaluating decontamination methods in ICU patients, SOD reduced mortality compared with standard of care (no decontamination) or placebo (OR 0.85, 95% CI 0.74-0.97) [124]. SDD also reduced mortality compared with standard of care (no decontamination) or placebo (OR 0.73, 95% CI 0.64-0.84).

●About half of the patients in the above meta-analysis were from a cluster-randomized trial of mechanically ventilated ICU patients in the Netherlands, in which SOD was performed with topical tobramycin, colistin, and amphotericin B, and SDD was performed with four days of intravenous cefotaxime in addition to tobramycin, colistin, and amphotericin B administered topically and through a nasogastric tube [117]. The interventions did not substantially reduce crude 28-day mortality rates (26.6, 26.9, and 27.5 percent with SOD, SDD, and standard care, respectively) but were associated with a lower odds of mortality compared with standard care in adjusted analyses. In a subsequent analysis of that trial, SOD and SDD were also associated with decreased rates of bacteremia and respiratory tract colonization with highly resistant bacteria [125].

●In a randomized trial including more than 11,900 ICU patients in the Netherlands designed to evaluate the rates of antimicrobial resistance with SOD versus SDD, mortality rates with the two interventions were similar [126]. Rates of rectal colonization with highly resistant bacteria were overall lower with SDD than SOD, despite a slightly greater increase over time in colonization with aminoglycoside resistant gram-negative bacilli with SDD than SOD. There was no comparison group in which no selective decontamination was used, so the resistance rates without SOD or SDD are unknown. In addition, these results may not be generalizable to locations in which the burden of antibiotic resistance is greater than that in the Netherlands.

●In a subsequent randomized trial including more than 8600 patients in 13 ICUs (in Belgium, Spain, Portugal, Italy, Slovenia, and the United Kingdom) with a moderate to high prevalence of antibiotic resistance (where at least 5 percent of bloodstream infections were caused by extended-spectrum beta-lactamase-producing Enterobacteriaceae) use of chlorhexidine mouthwash, SOD or SDD was not associated with reductions in mortality or in ICU-acquired bloodstream infections caused by multidrug-resistant gram-negative bacteria compared with standard care (which included chlorhexidine bathing and a hand hygiene improvement program) [119].

Issues related to the impact of SOD, SDD, and oropharyngeal chlorhexidine on rates of ventilator-associated pneumonia are discussed separately. (See "Risk factors and prevention of hospital-acquired and ventilator-associated pneumonia in adults", section on 'Decontamination of the oropharynx and digestive tract'.)

Surveillance — Surveillance for infections with multidrug-resistant bacteria within the institution as a whole and within specific units is critical for the early identification and control of epidemic outbreaks and endemic increases of resistant bacteria. The incidence and prevalence of isolation of multidrug-resistant bacteria (eg, MRSA, VRE, and carbapenem-resistant Enterobacteriaceae) should be monitored, and these data should be disseminated to nurses and clinicians who work in the ICU through a form that is easy to interpret. It is useful to compare data from different time periods for one ICU and to compare different units within the same hospital. The United States National Healthcare Safety Network System also provides comparisons of rates among participating hospitals [6]. Comparing rates among different institutions helps to identify hospitals or units where problems persist and can help to gauge the efficacy of interventions aimed at controlling rates of endemic and epidemic resistant organisms. (See "Vancomycin-resistant enterococci: Epidemiology, prevention, and control", section on 'Surveillance cultures' and "Methicillin-resistant Staphylococcus aureus (MRSA) in adults: Prevention and control".)

Positive surveillance cultures results should be coupled with effective flagging systems and established routes of communication inside the facility and between neighboring facilities in order to be successful [127]. In theory, this measure can reduce patient-to-patient transmission rates.

Active surveillance cultures, or screening patients for asymptomatic colonization with resistant organisms, is widely performed but may not be as effective as universal decolonization at controlling the spread of certain resistant pathogens in ICU settings. This is illustrated by the large multicenter trial described above in which universal chlorhexidine bathing and nasal mupirocin reduced rates of MRSA clinical isolates and bloodstream infections from any pathogen to a greater extent than did screening for MRSA and isolating carriers [77] (see 'Decolonization/patient bathing' above). This study did not evaluate active screening for other drug-resistant organisms, such as VRE. (See "Vancomycin-resistant enterococci: Epidemiology, prevention, and control", section on 'Surveillance cultures'.)

Prior to this study, the efficacy of universal active surveillance cultures for MRSA was uncertain, as illustrated by the following studies with conflicting conclusions [128-131]:

●In a cluster-randomized trial involving more than 9000 patients admitted to 18 intensive care units, use of MRSA surveillance cultures and expanded use of barrier precautions (universal glove precautions while awaiting active surveillance culture results) were not effective in reducing transmission of MRSA compared with existing practice [130]. This finding was surprising given that surveillance cultures identified a sizable subgroup of colonized patients who were not otherwise recognized. The authors hypothesized that additional interventions such as antiseptic bathing and improved environmental decontamination may be needed. Limitations of the study included the long turn-around time of the active surveillance cultures and omission of gowns as part of the barrier precautions.

●In a Veterans Affairs (VA) system-wide quality improvement initiative including nearly 2 million patients in 150 hospitals, a MRSA "bundle" program (including universal MRSA surveillance, contact precautions for colonized or infected patients, hand hygiene, and institutional culture change) was implemented [131]. This program was temporally associated with a reduction in the rate of health care-associated MRSA infection in intensive care units by 62 percent (from 1.64 to 0.62 infections per 1000 patient-days) and general units by 45 percent (from 0.47 to 0.26 infections per 1000 patient-days). However, as this study did not include control groups, it was not possible to determine whether MRSA surveillance was causally related to the observed drop in rates.

●A mathematical model demonstrated that the universal screen and isolate strategy could have contributed only marginally to the reduction in infections, since transmission rates before bundle implementation were already low and most patients with MRSA colonization were already colonized at admission [132].

However, there are other pathogens (such as VRE and carbapenem-resistant Enterobacteriaceae) for which active surveillance may continue to be an important measure in the efforts to contain the spread of these organisms in health care settings [90]. Moreover, since decolonization may not be practical or effective for these pathogens, active surveillance and strict contact isolation for carriers is still important for limiting spread [133]. Screening for asymptomatic carriage of C. difficile is gaining interest but data are not yet conclusive [134]. There are no reliable surveillance techniques to detect carriage of P. aeruginosa and/or A. baumannii [135]. (See "Vancomycin-resistant enterococci: Epidemiology, prevention, and control", section on 'Surveillance cultures'.)

Device-specific strategies — Preventing infections and decreasing the length of hospital stay of patients will decrease antimicrobial utilization and decrease the risk of becoming infected or colonized with resistant bacteria. Limiting unnecessary use of central venous catheter, bladder catheter, and endotracheal intubation decreases infection rates, decreases antibiotic use, and decreases selective antibiotic pressure on resident bacteria. Clinicians should assess on a daily basis the need to keep each of these invasive devices in place [136].

In addition, as many of the multidrug-resistant infections in the ICU are associated with indwelling devices, specific strategies for placement and care of such devices, as well as additional adjunctive measures, are effective in reducing the risk of catheter-associated urinary tract infections, ventilator-associated pneumonia (table 1), and intravascular catheter-related bloodstream infection (table 2). These strategies are discussed in detail elsewhere:

●(See "Placement and management of urinary bladder catheters in adults" and "Complications of urinary bladder catheters and preventive strategies", section on 'Prevention of complications'.)

●(See "Risk factors and prevention of hospital-acquired and ventilator-associated pneumonia in adults", section on 'Prevention'.)

Environmental cleaning — Environmental cleaning, disinfection, and sterilization are basic and important measures used to prevent or reduce infections in the intensive care unit, as in the rest of the hospital environment. Innovative but as yet experimental techniques for environmental cleansing include ultraviolet light sterilization lamps and hydrogen-peroxide vapor decontamination devices, which might contribute to future attempts at reducing colonization pressure. However, these new technologies will not replace the necessity of proper manual "terminal" cleaning that should be established by a written protocol in every ICU, and adherence to the protocol must be regularly monitored. (See "Infection prevention: General principles", section on 'Health care environment: Cleaning and disinfection'.)

SUMMARY AND RECOMMENDATIONS

●Infections are especially frequent in intensive care units (ICUs) because the patients are likely to have chronic illnesses and acute physiologic derangements, indwelling catheterization is common, and multidrug-resistant pathogens that are difficult to eradicate are isolated with increasing frequency due to enhanced selective antimicrobial pressure and enhanced colonization pressure. (See 'Introduction' above and 'Epidemiology' above.)

●The most common multi- and extensively drug-resistant pathogens isolated in ICUs include methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), carbapenem-resistant Pseudomonas aeruginosa, Acinetobacter baumannii, and extended-spectrum beta-lactamase-producing Enterobacteriaceae. Carbapenem-resistant Enterobacteriaceae are also increasing in prevalence. (See 'Risk factors for resistant infections' above.)

●Comorbid conditions, long hospital courses, frequent contact with health care personnel, indwelling catheterization, and receipt of antimicrobial therapy all increase the risk of colonization and infection with multidrug-resistant pathogens. Infections with such organisms are associated with increased mortality, length of stay, and hospital costs. (See 'Risk factors for resistant infections' above and 'Outcome of multidrug-resistant infections' above.)

●Good hand hygiene compliance, contact precautions for patients who harbor epidemiologically relevant drug-resistant organisms, and minimizing unnecessary hospitalization and interventions are critical for preventing infection and the spread of resistant organisms in the ICU. Adequate and standardized approaches to environmental cleaning and disinfection is an additional established measure to contain the spread of multidrug-resistant organisms. More intensive infection control interventions to reduce colonization pressure include cohorting with dedicated staff, chlorhexidine bathing, selective decontamination, active surveillance for certain pathogens, and reduction of catheterization utilization. (See 'Infection control measures' above and "Infection prevention: Precautions for preventing transmission of infection".)

●Restricted and judicious antibiotic utilization, often implemented as part of a global institutional antimicrobial stewardship program, can decrease selective pressure that promotes emergence of resistant bacterial strains. Infection control measures such as hand hygiene prevent the spread of multidrug-resistant organisms. (See 'Prevention' above.)

●For all patients in an ICU, we suggest daily chlorhexidine bathing (Grade 2C). Daily chlorhexidine bathing decreases the risk of colonization and infection with drug-resistant and other organisms and is associated with minimal adverse effects. Additionally, it may be more effective in controlling certain infections than an active surveillance policy. (See 'Prevention' above.)

●The most common infections in the ICU are those associated with indwelling devices, namely catheter-associated urinary tract infection, ventilator-associated pneumonia, and intravascular catheter-related bloodstream infection. Apart from minimizing their use, certain strategies regarding placement and care of indwelling devices can decrease the risk of infection (table 1 and table 2). The epidemiology, management, and prevention of these infections are discussed in detail elsewhere.

•(See "Catheter-associated urinary tract infection in adults" and "Complications of urinary bladder catheters and preventive strategies".)

•(See "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults" and "Treatment of hospital-acquired and ventilator-associated pneumonia in adults" and "Risk factors and prevention of hospital-acquired and ventilator-associated pneumonia in adults".)

(See "Intravascular catheter-related infection: Epidemiology, pathogenesis, and microbiology" and "Intravascular non-hemodialysis catheter-related infection: Treatment".)