International conference on health care–associated bloodstream infections

Health care-associated blood stream infections, November 8-9, 2007 

Health care-associated infections (HAIs) are common complications in hospitals throughout the world. In the United States alone, the most recent study estimated that 1.7 million HAIs occur annually and result in 99,000 deaths. The number of deaths associated with HAIs was greatest for pneumonia and catheter-related blood stream infections (CR-BSIs).1 The epidemiology of HAIs has become more complex because health care is being provided outside acute care hospitals in settings such as long-term acute care, outpatient facilities, and the medical home environment. Hospital pathogens are being transmitted both within these settings and after discharge in the community.

Despite growing evidence that hospitals can significantly reduce HAIs, until recently infections were considered an inevitable part of care. The key to success has been to combine evidenced-based interventions (bundles) with the adaptive cultural changes necessary to achieve actionable, sustainable changes.

The importance of changing organizational culture to enhance safety was the key message presented by Peter Pronovost to open the conference. Dr. Pronovost discussed the Keystone Project in Michigan that involves most intensive care units (ICUs) in the state. The results showing the impact on preventing CR-BSIs were published in 2006. A large and sustained reduction of infections was demonstrated from a CR-BSI medium rate of 2.7/1000 catheter-days at baseline down to 0/1000 catheter-days at 18 months.2 The Institute for Healthcare Improvement's To Save 100,000 Lives Campaign also included interventions to reduce CR-BSIs and resulted in a decreased incidence of deaths because of CR-BSIs. The Voluntary Hospitals of America's Transformation of the ICU project demonstrated similar reductions in CR-BSIs using a similar approach of combining a catheter bundle and adaptive changes. Most catheter bundles include handwashing, maximum barrier precautions during line insertion, chlorhexidine skin preparation, avoidance of femoral site if possible, and removal of unnecessary catheters.

An estimated 68.8% of HAIs occur outside the ICUs and often are missed by traditional focus surveillance.1 This was highlighted by the presentation and publication by Jonas Marschall, MD, on CR-BSIs in general medicine patients not receiving care in an ICU.3 They found that non-ICU patients had lower device utilization rates but similar CR-BSIs rates compared with medical ICUs. This study suggested that the same HAI preventive strategies should be implemented outside the ICU.

Carlene Muto, MD, discussed the efforts of Southwestern Pennsylvania and the University of Pittsburgh Medical Center in reducing CR-BSIs. The initial findings were published in 2005,4 and Dr. Muto updated these remarkable results to include data from 2007. She emphasized that CR-BSIs are not inevitable but can be substantially reduced by the implementation of standardized targeted practices that she called “bundles of joy.” Results showed that 158 CR-BSIs were avoided annually with a cost avoidance of $5.2 million.

Diagnosis of a CR-BSI can be challenging and often requires catheter withdrawal. Emilio Bouza, MD, PhD, conducted a randomized prospective study of 3 techniques for the diagnosis of CR-BSIs without catheter withdrawal. His team compared semiquantitative cultures from skin and catheter hub (considered positive when there were ≥15 colonies of the same organism isolated from a skin and hub sample as from a peripheral blood culture), differential quantitative blood cultures, and differential time to positivity. Surprisingly, all 3 techniques had comparable accuracy. Based on ease of performance and cost, semiquantitative skin and hub cultures in combination with conventional peripheral blood culture are recommended, with the other technique reserved as a confirmatory technique in difficult cases.5

Richard Johannes, MD, examined the epidemiology of early BSIs (≤5 days from admission). The initial results were published in 2006.6 The investigators studied 3 categories: hospital-acquired bloodstream infection (HAB), health care-associated bloodstream infection (HCAB), and community-acquired bloodstream infection (CAB). HCAB has traditionally been classified within the CAB group. Using this new classification, HCAB was found to be clinically closer to HAB than CAB, and, therefore, treatment should be similar to HAB in the absence of culture data. Multivariable logistic regression analysis showed that the mortality for HCAB was intermediate to that of CAB and HAB. More recently, as reported in proceedings and at the 46th Interscience Conference on Antimicrobial Agents and Chemotherapy meeting,7 the significance of systemic inflammatory response syndrome and organ dysfunction in culture-positive bacteremic patients was examined. Evidence of one or more major organ dysfunction with or without systemic inflammatory response syndrome was associated with higher mortality rates.

Michael Smith, MD, discussed the unique characteristics of children that place them at risk for CR-BSIs. He concluded that, despite limited data for specific interventions, a multifaceted intervention approach has been effective in reducing CR-BSIs.8 He also presented data about the efficacy of antimicrobial lock therapy in certain situations.9 Finally, he discussed the use of a maintenance bundle recently implemented in the 26-bed pediatric ICU at the University of Louisville. This bundle consisted of daily assessment of the need for a line, use of cap change kits, hand hygiene and glove use with line access, a 15-second alcohol scrub with any line entry, and use of chlorhexidine patches in all patients. Initial results indicate a dramatic decrease in CR-BSIs.

Juan Yébenes, MD, PhD, discussed the unintended consequences of adopting needle-free devices to prevent bloodborne pathogens via needlestick injuries. Although needlestick injuries have declined, published results show an increase in CR-BSIs related to poor compliance with safe practices.10 Dr. Yébenes considered the importance of the barrier effect of these needle-free devices related to external contamination and the use of the correct disinfection before handling. He emphasized the importance of future device design in improving the barrier effect. In another presentation, Steve Menyhay, RN, presented a novel antiseptic-barrier cap that prevented entry of bacteria even with heavy contamination of the septum of a needle-free connector. The next step will be to evaluate this technology with CR-BSIs as the primary outcome measure.

Victor Rosenthal, MD, presented the impressive findings of the International Nosocomial Infection Control Consortium's efforts on reducing device-related infections in developing countries, where rates of HAIs and bacterial resistance are 3 to 5 times higher than in the United States.11 Through process surveillance, education, and performance feedback, the International Nosocomial Infection Control Consortium was able to show significant increases in handwashing compliance and other infection control interventions. More importantly, these techniques resulted in significant decreases in mortality and overall rates of HAIs and CR-BSIs.

Peripherally inserted central venous catheters (PICCs) are widely used for outpatient and inpatient intravenous therapy. Basel Al Raiy, MD, presented the results of a prospective study of the risk of PICC-related infectious and thrombotic complications (deep vein thrombosis) in hospitalized patients with PICCs. The majority of PICCs were double-lumen, with 60% placed by an IV team and 40% by interventional radiology. Results showed a CR-BSI rate of 3/1000 catheter-days if the PICC was placed by an intravenous therapy team compared with 0.97/1000 catheter-days if placed by interventional radiology. Multivariate analysis showed that 2 factors were significant for infection: duration of line use and female sex. Deep vein thrombosis was very uncommon in those with PICC placement.

Marianne Opilla, RN, reviewed the epidemiology of BSIs associated with parenteral nutrition (PN). The PN composition of amino acids, dextrose, and fat emulsions supports the growth of fungi and bacteria. In long-term PN patients, common BSI pathogens include Staphylococcus species, gram-negative bacteria, and Candida species.12 Several nutritional strategies have been suggested to prevent BSIs in patients receiving PN. Continued need for PN should be assessed daily. The enteral route is preferred over parenteral nutrition care because of lower complication rates. Tight glycemic support has been shown to reduce morbidity and mortality in ICU patients.13 Immunonutrition with use of omega-3 and glutamine supplements has shown some promise but needs further study. The use of catheter lock therapy has been shown to reduce levels of intraluminal biofilm in catheters in place for more than 2 weeks.14 Ethanol lock therapy also was shown to reduce the incidence of CR-BSIs in home PN patients.15

Saima Aslam, MD, reviewed the role of biofilm in the pathogenesis of CR-BSIs and discussed ways to reduced biofilm-associated infections: prevent adherence of microorganisms by using antibiotic- or silver-impregnated catheters; penetrate the biofilm with high-dose antibiotics, and/or ethanol delivered by catheter lock; use quorum-sensing inhibitory compounds; and attack the structural integrity of a biofilm with enzymes or n-acetylcysteine with or without antibiotics. In a recent in vitro study, Dr. Aslam et al demonstrated that the combination of NAC-tigecycline significantly decreased viable biofilm bacteria.16 A pilot clinical trial is planned.

Many studies have demonstrated increased morbidity and mortality associated with CR-BSIs. The true cost of BSIs is less clear, based on studies using different methodologies. Meredith Kilgore, RN, PhD, and Steve Brossette, MD, PhD, presented a model that controlled for length of stay and the presence of other HAIs. Costs were obtained by matching laboratory data with hospital accounting system calculations of both fixed and variable costs of care. The basic regression analysis estimated the cost of a BSI to be $19,643. If the analysis controlled for the presence of other HAIs, then BSI was associated with an incremental cost of $12,774. Finally, if the analysis controlled for length of stay, the estimated cost was $5534. Electronic surveillance allowed this study to be conducted at very low cost.

Discussion of antimicrobial resistance is essential to any consideration of HAIs. The United States and the rest of the world are currently in the throes of an epidemic of antimicrobial-resistant infections caused by multidrug resistant organisms. These include methicillin-resistant Staphylococcus aureus, vancomycin-resistant enterococcus (VRE), multidrug-resistant gram-negative bacilli, Mycobacterium tuberculosis, and Candida species. Many studies have demonstrated increased morbidity, mortality, and cost associated with multidrug-resistant organisms. Inappropriate use of antibiotics increases selective pressure that increases resistance. Inadequate infection prevention and control promotes transmission of multidrug-resistant organisms, further complicating the problem.17, 18

Cassandra Salgado, MD, discussed the epidemiology and role of VRE colonization and infection in hospitals. In 2003, Drs. Salgado and Farr published a metaanalysis of outcomes associated with VRE.19 Results showed that compared with vancomycin-sensitive enterococcus, VRE infections were associated with higher rates of recurrent BSIs, higher case-fatality rates, and higher costs. At the conference, Dr. Salgado presented data showing that patients with VRE tended to be sicker longer, had increased lengths of stay, and higher mortality rates. VRE colonization was a precursor to VRE-BSIs. Results showed that control and prevention of VRE can reduce morbidity, mortality, and health care costs. Recommendations to control VRE-HAIs include antimicrobial stewardship; reducing patient-to-patient transmission by reducing environmental contamination; reducing health care worker contamination through the use of hand hygiene, gowns, and gloves; and using the CR-BSI bundle to prevent bacteremias.

Kevin Garey, PharmD, discussed work on the importance of early treatment of candidemia in reducing mortality. Candida species are now the fourth leading cause of health care-associated BSIs.20 Inadequate antimicrobial therapy has been shown to be an independent determinant of mortality related to health care-associated infections, including fungal BSIs.21 We have also seen a trend toward candidemia caused by less susceptible Candida species, including Candida glabrata and Candida krusei. Dr. Garey showed that a delay in the initiation of fluconazole therapy in hospitalized patients with candidemia increased mortality.22 He also discussed the importance of increasing fluconazole doses to achieve an adequate dose/minimum inhibitory concentration ratio.

Finally, Jeanne Zack, RN, PhD(c), discussed the cultural change at Missouri Baptist Medical Center to zero tolerance for infection and the important role of administration and the Board. This was an important theme throughout the conference. Zero tolerance does not mean zero infections but rather rejecting the notion that HAIs, including CR-BSIs, are inevitable. The essential elements for a change to zero tolerance and reducing the rate of HAIs include evidenced-based practice, measurement and feedback, collaboration and teamwork at all levels, leadership support, continuous learning, and everyone being accountable for compliance.

In summary, we have learned that no single intervention prevents an HAI. A “bundle” of evidenced-based interventions implemented by multidisciplinary teams can produce greater and more sustained results. Second, health care is moving from benchmarking infections to a culture of zero tolerance. Last, a culture of accountability and transparency requires active administrative and Board support to provide care safely and reliably to our primary customer, our patients.