The Commonality of Risk Factors for Nosocomial Colonization and Infection with Antimicrobial-Resistant Staphylococcus aureus, Enterococcus, Gram-Negative Bacilli, Clostridium difficile, and Candida

Beginning in the late 20th century, we have witnessed a rapidly growing crisis in antibiotic resistance, especially among microorganisms that cause nosocomial infection (1-4). Most notable among these are methicillin-resistant Staphylococcus aureus (5), vancomycin-resistant enterococcus (6), Clostridium difficile (7), extended-spectrum -lactamase-producing gram-negative bacilli (8), and Candida (9). Infections caused by these microorganisms increase hospital stays and attributable mortality (1-9). The success of programs to curtail antimicrobial resistanceespecially in controlling endemic infectionshas been limited. Strategies for preventing nosocomial infection are more likely to succeed if they are guided by a full understanding of the factors that put hospitalized patients at increased risk (10). Numerous studies have identified risk factors for nosocomial colonization or infection by individual multiresistant pathogens. After reviewing the published studies on this subject, we conclude that the most important risk factors for colonization or infection with the various multiresistant microorganisms are common and universal. In this review, we do not differentiate between colonization or infection because nosocomial colonization is the precursor to clinical infection. The risk for nosocomial infection is 11% to 38% with colonization by methicillin-resistant S. aureus (11, 12), 25% with colonization by vancomycin-resistant enterococcus (13), 25% with colonization by extended-spectrum -lactamase-producing gram-negative bacilli (14), and as high as 38% with colonization by Candida (15). We included in our review published papers or abstracts that used multivariable techniques of statistical analysis to identify risk factors in adults. We found 74 studies with data on nosocomial colonization (n = 24) or infection (n = 50) by methicillin-resistant S. aureus (n = 13) (11, 12, 16-26), vancomycin-resistant enterococcus (n = 22) (27-48), extended-spectrum -lactamase-producing gram-negative bacilli (n = 9) (49-57), C. difficile (n = 20) (58-77), or Candida (n = 10) (78-87). Only 21 studies (29%) were conducted prospectively; the remainder retrospectively analyzed colonization or infection as determined by clinical cultures obtained by patients' caregivers. Most studies examined few risk factors; only 59 (80%) quantified risk by calculating odds ratios or relative risk. Thirty-five studies (49%) examined a broad, general hospital population; the remainder focused on selected subgroups, such as patients in an intensive care unit (ICU) (16 studies), patients with HIV infection or malignant solid organ or hematologic conditions (7 studies), patients in nursing homes (6 studies), patients receiving hepatic transplants (4 studies), patients undergoing general surgery (3 studies), or patients requiring hemodialysis (2 studies). Table 1 shows risk factors that predicted nosocomial colonization or infection with individual multiresistant organisms. Seven types of risk factors were most likely to result in colonization or infection with multiresistant species: advanced age (odds ratio [OR], 1.2 to 14.1); severity of illness (OR, 1.9 to 11.6); inter-institutional transfer of the patient (OR, 2.9 to 21.3); prolonged hospital stay (OR, 1.3 to 17.5); gastrointestinal surgery (OR, 2.5 to 6.9); transplantation (OR, 3.2 to 6.7); exposure to medical devices, especially central venous catheters (OR, 1.8 to 26.4); and, universally, heavy exposure to broad-spectrum antimicrobial drugs (OR, 1.6 to 25.1), especially cephalosporins (OR, 1.6 to 28.6). Table 1. Risk Factors for Nosocomial Colonization or Infection with Methicillin-ResistantStaphylococcus aureus, Vancomycin-ResistantEnterococcus, Clostridium difficile, Extended-Spectrum -Lactamase-Producing Gram-Negative Bacilli, and Candida Discussion Numerous reports have documented a striking increase in the incidence of nosocomial infection caused by the multiresistant species examined in this study (1-9). Data on ICU patients from the nearly 300 U.S. hospitals participating in the National Nosocomial Infection Surveillance Study of the Centers for Disease Control and Prevention (88) showed that by 1998, nosocomial infections caused by methicillin-resistant S. aureus accounted for 55% of all S. aureus infections; those caused by vancomycin-resistant enterococcus, 26% of all enterococcal infections; and those caused by gram-negative bacilli resistant to third-generation cephalosporins (and presumably, extended-spectrum -lactamase producers), 36% of all Enterobacter infections, 20% of all Psuedomonas aeruginosa infections, and 9% of all Klebsiella infections. A similar prospective study of antimicrobial resistance in 40 U.S. medical centers, the Intensive Care Antimicrobial Resistance Epidemiology Project (89), documented similar trends. The five pathogens we reviewed now account for approximately 50% of all nosocomial infections in U.S. hospitals (88, 89). These trends in endemic nosocomial infection, as well as the increasing number of reports of outbreaks of Candida infection (90), indicate that current infection-control efforts to target multiresistant pathogens are failing dismally. Although many studies have identified risk factors that are specific to individual pathogens, few studies have attempted to establish commonality of risk factors among these organisms or even to explore their relationships with each other. It has been shown that infection by C. difficile can facilitate transmission of vancomycin-resistant enterococcus (43, 91) and that successful control of vancomycin-resistant enterococci results in a commensurate reduction in C. difficile infections (92). In a review of the relationship between vancomycin-resistant enterococcus and C. difficile, Gerding (93) showed that persons infected with these organisms share many common risk factors, including type of antimicrobials received; advanced patient age; extended length of stay; severity of underlying disease; and exposure to electronic thermometers, enteral feeding, and environmental contamination. Our analysis is limited by the modest size of the included studies and, especially, by their mostly retrospective nature, which limits confidence that all of the control patients were indeed free of colonization by the resistant pathogen in question. Nonetheless, our analysis shows an impressive commonality of risk factors across the five groups of diverse multiresistant pathogens. We believe these data indicate that infection-control programs that focus on one organism or only one antimicrobial agent are unlikely to succeed. For maximum benefit, we believe that infection-control programs must apply global strategies aimed at all resistant organisms (Table 2)(94-153). Table 2. Risk Factors and Potential Control Measures for Colonization or Infection with Antimicrobial-Resistant Nosocomial Pathogens Amenable to Control Severity of Illness and Underlying Disease and Extended Length of Stay It is clear that patients with serious underlying disease and a high severity-of-illness score, particularly those who have undergone complicated surgery or organ transplantation or who have renal or other organ failure, are at greatly increased risk for infection by multiresistant organisms. This increase in risk results from the patients' exposure to invasive devices and other procedures (each of which increases risk), as well as from depressed host defenses and extensive exposure to antibiotics. Almost invariably, the duration of hospitalization is considerably longer in these patients than in patients with less severe illness. Obviously, little can be done to modify illness severity. Thus, we believe that strategies to protect intrinsically vulnerable patients (Table 2) must focus on limiting and improving the use of antimicrobial therapy; more consistent application of basic infection-control measures, such as hand-related hygiene; and the adoption of novel prevention-related technology, such as anti-infective-coated vascular (95-97) or urinary (98-101) catheters. However, we should also consider the reality that our current paradigm for preventing spread of resistant organisms in the hospitalwaiting until colonization or infection by a resistant organism is reported by the laboratory before isolating the patient, usually in a single room, and requiring gloves with or without a gown for all contacts with that patienthas failed. Prolonged wearing of gloves in the hospital, a very common practice, may paradoxically increase the risk for nosocomial cross-infection (154, 155). Considerable data indicate that cohort nursing for patients with known colonization or infection by multiresistant organisms can reduce cross infection (102-105). However, for every patient with known colonization of infection by multiresistant organisms, many more patients with unrecognized colonization are already in the medical centerprobably in the same unit (20, 33, 43, 44, 156). A more logical strategy is the preemptive use of barrier precautions for all high-risk patients to prevent contamination of the hands of health care workers by multiresistant organisms during contact with patients who have silent colonization and to prevent cross-infection to other uncolonized patients. Studies have shown that the preemptive use of barrier precautionsalso called protective isolationis highly effective at preventing the spread of multiresistant organisms, such as methicillin-resistant S. aureus or vancomycin-resistant enterococcus in an epidemic setting (107-109). Some studies have also shown a strong correlation between preemptive use of barrier precautions in high-risk populations, such as patients in the intensive care unit (ICU), and prevention of endemic nosocomial infection, including by multiresistant organisms (110-114). Objects in the hospital environment also appear to play a significant role in the transmission of resistant pathogens. The use of stethoscopes (157),