Sharon F. Welbel
Robert A. Weinstein
Technical, computer, and radiologic advances over the past 10–20 years have facilitated dramatic growth in the fields of diagnostic and therapeutic procedures. With miles of vascular network, dozens of extravascular spaces, and several organ systems, any patient is a candidate for a staggering array of such interventions. Although many of these procedures can provide information that is essential for sophisticated patient care, supplant more traumatic interventions, and are critical for life support, most procedures also bypass natural host defenses and place patients at increased risk of healthcare-associated infection (HAI) [1,2]. It is not surprising, then, that the introduction of any new procedure often is followed closely by case-reports of procedure-associated infections. Occasionally, epidemiologic experiments of nature in the form of HAI outbreaks provide more detailed information on specific procedure-related hazards, which eventually could be subjected to prospective study. This chapter discusses a variety of procedure-associated infections that have been highlighted by retrospective or prospective investigations and that have not been discussed elsewhere in this book.
Because of the seemingly eclectic contents of this chapter, it is important to recognize from the outset that the procedures to be discussed have certain themes in common. First, all of the procedures are exquisitely vulnerable to these possibilities: inexperienced operators; breaks in aseptic technique; contaminated, inadequately disinfected, and/or technically difficult-to-clean equipment; and ineffective antiseptics. Second, many procedure-related infection problems unfortunately reemerge as new generations of healthcare workers (HCWs) rediscover these vulnerabilities and/or are being newly recognized in developing countries that have fewer resources to devote to infection control, hence the importance of reviewing hazards that at first glance could appear remote [3]. Third, various procedures involving many different sites have as a common path of infection the bloodstream, although the risk of infection differs depending on whether bloodstream contamination is transient or persistent as well as on host- and organism-specific factors. Fourth, risks of biofilm formation on surfaces of many different devices that enter normally sterile body sites and the value of anti-infective coatings (e.g., antimicrobials, antiseptics, and heavy or precious metals) for such devices are being studied actively. Finally, many procedures bear the burden that the specific risks have not been defined sufficiently to determine whether certain preventive measures, such as the use of prophylactic antimicrobial therapy, are mandated.
Infections from Procedures Involving the Vascular System
Needleless Devices
HCWs always have been at risk of needlestick injuries, but quantifying the actual number of percutaneous needlestick
P.778
injuries (NSIs) sustained by U.S. HCWs is difficult. Panlilio et al. combined data for 1997 and 1998 from the National Surveillance System for Healthcare Workers (NaSH) and the Exposure Prevention Information Network (EPINet) and adjusted the data for underreporting. The investigators estimated that the number of percutaneous NSIs sustained annually by U.S. HCWs is 304,325 [4]. This is in contrast to the 1,728 percutaneous NSIs reported in 2003 by 48 U.S. healthcare facilities to the EPINet surveillance program. The overall annual percutaneous NSI rate for all network hospitals in 2003 was 23.87 per 100 occupied beds [5]. Clearly, the risk of transmission of blood borne pathogens, such as human immunodeficiency virus (HIV), Hepatitis B virus (HBV), and Hepatitis C virus (HCV), still exists. The current federal standard for addressing NSIs among HCWs is the blood-borne pathogens standard promulgated by the Occupational Safety and Health Administration (OSHA) [6]. This standard requires that engineering controls and work practices eliminate or minimize HCW exposure to blood. One means of accomplishing this goal is to use needleless systems.
Since the introduction of needleless systems, decreased rates of occupational needlestick exposures have been documented [7,8,9]. Protected-needle intravenous (IV) systems also have decreased NSIs related to IV connectors by 62–88% [10,11]. Unfortunately, the devices are not routinely activated, which appears to be related to inadequate training [9].
The impact for the patient is less clear. A study to determine risk factors for bloodstream infections (BSIs) in patients receiving home intravenous infusion therapy [12] revealed that receipt of total parenteral nutrition and intralipid therapy through a needleless system was a BSI risk factor. The results of a survey on the subject of injection caps demonstrated that positive cultures were significantly more common from needleless devices than from protected-needle devices. It was concluded that when injection caps are manipulated, nutrient-rich solutions can remain in the caps of the needleless devices and become contaminated. Another study that assessed needleless systems used with Hickman catheters suggested that such systems can be associated with increased rates of catheter-related BSI [13]. The investigators cultured luminal fluid from Hickman catheters of hematology patients and found that these catheters with the needleless system were twice as likely to show luminal contamination compared with catheters without the system. Four BSIs in patients with the needleless device had peripheral blood and luminal fluid cultures that yielded concordant bacterial strains based on results of pulsed-field electrophoresis and restriction fragment polymorphism studies. Do et al. described an increased BSI rate with the use of a needleless intravenous system in a home care setting when caps were changed every 7 days and a subsequent decrease in BSIs when the needleless device end cap was changed every 2 days, suggesting that the mechanism for BSI could involve contamination from the end cap [14]. Kellerman et al. reported an 80% increase in BSIs related to central venous catheters (CVC) in pediatric hematology oncology patients receiving home health care after introduction of a needleless device for CVC access [15]. At another institution, a significant increase in the BSI rate in a surgical intensive care unit (ICU) and an organ transplant unit was associated with the introduction of a needleless intravenous system. This was attributed to nurses' lack of familiarity with the device and deviation from the manufacturers' recommended practices [16]. Finally, another study that investigated risk factors associated with an increased rate of BSIs in pediatric ICU found that exposure to the IVAC first-generation needleless device (IVAC, San Diego, California) was an independent BSI risk factor. The BSI rate returned to baseline after institution of a policy to replace the entire IVAC device, valve, and end-cap every 24 hours [17].
The association between needleless devices and infection seems to relate to lack of familiarity with the device and/or its mechanics. Some investigators have investigated the mechanics of needleless devices to determine whether new technology added to the device could reduce infection risk. Menyhay and Maki reported on a simulation study that compared the efficacy of conventional alcohol disinfection of the membranous septum of needleless luer-activated valved connectors before access with the use of a novel antiseptic-barrier cap that, when threaded onto the connector, places a chlorhexidine-impregnated sponge in continuous contact with the membranous surface [18]. After removal of the cap, there is no need to disinfect the surface with alcohol before accessing it. After contaminating, disinfecting, and then culturing the connectors, the authors demonstrated that if the membranous septum of a needleless luer-activated connector is heavily contaminated, conventional disinfection with 70% alcohol did not reliably prevent entry of microorganisms. In contrast, the antiseptic-barrier cap provided a high level of protection. Another study considered a recently developed needleless closed luer access device (CLAD) (Q-Syte, Becton Dickinson, Sandy, Utah). Devices were contaminated with bacteria and then disinfected with 70% isopropyl alcohol followed by flushing with 0.9% saline. Although devices had been accessed up to 70 times, no microorganisms were found even when challenge microorganisms were detected on the syringe tip after activation and on the compression seals before decontamination, suggesting that the Q-Syte CLAD can be activated up to 70 times with no increase risk of microbial contamination within the fluid pathway [19]. Needleless systems are now almost universally used; the benefit to the HCW and the risk to the patient have been demonstrated. Education is a key intervention to prevent patient infections with devices new to an institution. Novel interventions as mentioned previously will need to be studied further to assess their benefit and cost. The luer-activated valved connector, which allows a chlorhexidine-impregnated sponge to do the work, could
P.779
be particularly helpful because it does not depend on the action of an HCW for disinfection.
Finally, the tourniquet could function as a possible vehicle for cross-contamination of pathogens such as methicillin-resistant Staphylococcus aureus (MRSA). Leitch et al. examined the contamination rate of phlebotomy tourniquets with MRSA. The investigators found that the tourniquets were contaminated with MRSA 25% of the time; they believed that the contamination occurred via the phlebotomists' hands, not the patients' skin [20]. The practice for tourniquet use varies widely from single patient use to disposal upon the discretion of the phlebotomist. Given that multidrug-resistant organisms, such as community-acquired MRSA, now are ubiquitous, the need for practices such as discarding or disinfecting tourniquets with alcohol wipe between uses to prevent tourniquets from harboring pathogens must be considered and evaluated.
Leeches
Despite the popular appeal of highly sharpened, disposable phlebotomy needles for diagnostic bloodletting, leeches have, in fact, resurfaced as a specialized part of the reconstructive and microvascular surgeons' armamentarium (e.g., for salvage of congested flaps [21]). However, as with many other advances discussed in this chapter, leeches have an infectious risk [22]. Aeromonas hydrophila, normal gut flora of the leech, has caused wound infections in 2.4–20.0% of microsurgical procedures using leeches [23,24]. Aeromonassp. meningitis also has been associated with leech therapy [25]. In an attempt to decrease infectious complications of leeches, one group tried unsuccessfully to sterilize the gut of leeches [26]. Some investigators believe that aquariums filled with tap water to house leeches could contribute to the Aeromonas sp. problem [27]. Infection with Serratia marcescens also has been linked to leech therapy [28]. Understanding the nature of the leech's contamination (gut flora and surrounding environment) could help direct control efforts and prophylactic antimicrobial therapy.
Cardiac Catheterization
Serious local and systemic infections can result from cardiac catheterization procedures, particularly when contaminated instruments or ineffective antiseptics (e.g., dilute aqueous benzalkonium chloride) are used inadvertently or when breaks in technique occur in the cardiac catheterization laboratory. The major pathogens are staphylococci and gram-negative bacteria.
Up to 50% of patients undergoing cardiac catheterization could experience an increase in temperature of >1°C (1.8°F) within 24 hours after catheterization. Their fever, however, has been attributed to the use of angiocardiographic contrast material rather than to infection. In fact, bacterial endocarditis has been reported very rarely in large studies evaluating the complications of cardiac catheterization, and individual examples could have been due to initially undetected concurrent infection. The BSI rate after procedures in the cardiac catheter laboratory range from 0.11–18.00%. In a study of more than 22,000 patients undergoing invasive, nonsurgical, coronary procedures from 1991–1998, BSIs occurred in 0.11% at a median of 1.7 days after the procedure; in >4,000 patients undergoing coronary intervention, bacterial infections occurred in 0.64% and septic complications in 0.24% [29,30]. However, in 147 consecutive patients undergoing complex percutaneous coronary interventions (PCI), positive blood cultures were found in 18% immediately after the procedure and in 12% at 12 hours after the procedure [31].
Some studies reporting transient BSIs obtained blood cultures from the intravascular catheter or from the vessel from which the catheter had been removed. It is possible that some of the isolates represented contamination of the external part of the catheter or the site of insertion and that the incidence of BSI was actually less frequent. A study designed to assess this possibility obtained blood for culture by using standard techniques from a vein distant from the site of catheter manipulation [32]. Venous blood cultures of 106 patients, most whom had valvular heart disease, were obtained in this manner during cardiac catheterization, and all were sterile. Of 38 samples drawn through the catheter that was placed in the heart or aorta during the procedure, 3 grew diphtheroids or microaerophilic streptococci. The researchers concluded that the contamination of the hub end of the catheter with normal skin flora led to an overestimation of the BSI incidence. Removal of organisms by lung filtration also could have accounted, in part, for the failure to isolate organisms from distal sites. In either instance, it is clear that some contamination of the catheterization field had occurred.
Coronary stent placement, a newer procedure, now is routinely practiced yet has been linked to few reports of coronary stent infections. When such infections do occur, mortality and morbidity is high (Table 45-1). Once a stent has been placed, it is not removable; therefore, illuminating risk factors for stent infection is paramount [33,34,35,36,37].
Retrospective and prospective studies have illuminated various risk factors for PCI-associated BSI. These factors include difficult vascular access, multiple skin punctures, repeated catheterizations at the same vascular access site, extended procedure duration, use of multiple percutaneous transluminal coronary angioplasty (PTCA)-balloons, deferred removal of the arterial sheath, presence of congestive heart failure, and patient's age >60 years [29,30,38]. We should focus on nonpatient factors such as timely removal of the arterial sheath after percutaneous transluminal angioplasty to decrease HAI rates in the cardiac catheter laboratory. In addition, catheterization-associated infection should be infrequent with rigorous application of strict aseptic technique and adoption of the working principle that cardiac catheterization is a
P.780
P.781
surgical procedure. The Laboratory Performance Standards Committee of the Society for Cardiovascular Angiography and Interventions (SCAI) has published an updated guideline that addresses the increased utilization of the catheterization laboratory as an interventional suite with device implantation. The guide is divided into sections on the patient, laboratory personnel, and laboratory environment [39]. The guidelines can be accessed at www.scai.org.
|
TABLE 45-1 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Indwelling Arterial Catheters
Indwelling arterial catheters are used regularly in patients who require pressure monitoring or repeated blood gas determination (see Chapter 37). Although they provide information that is essential for patient care and eliminate the need for potentially traumatic repeated arterial punctures, such catheters also provide a continuing portal of entry for microbial invasion of the bloodstream. The reported incidence of arterial catheter colonization and infection varies depending on the catheter-tip culture technique used. Colonization incidence reports vary from 27% (49 episodes per 1,000 catheter-days) to 4% (11.7 episodes per 1,000 catheter-day) [40,41]. The source-organisms have not always been evaluated, and no direct relation with patient disease has been established, but the incidence of colonization of radial catheters (in contrast to umbilical catheters) did appear to be related to longer durations of catheterization (>4 days) [42]. Inflammation at the catheter site and the use of a cutdown procedure to place the catheter also appear to be associated with an increased infection risk [43,44,45,46] (see Chapter 37). A prospective study of 95 patients (130 catheters) in a medical–surgical ICU found a 4% risk of arterial cannula-related septicemia; 12% of all sepsis in this unit was the result of intra-arterial catheters [47]. These BSIs were caused by gram-negative bacilli, enterococci, or Candida organisms.
There are no national guideline recommendations for arterial catheter site insertion [48]. Intuitively, it seems that a femoral catheter would engender a greater risk of infection than a radial site, but few studies have demonstrated this. One study by Lorente et al. attempted to answer this question [49]; the authors performed a prospective observational study of all consecutive patients (2,018 patients with 2,049 arterial catheters) admitted to a medical–surgical ICU over 3 years. Multivariate analysis revealed that the catheter-related local infection rate was significantly higher for femoral than radial access (odds ratio [OR], 1.5; 95% confidence interval, 1.10–2.1). The catheter-related BSI rate also was higher (femoral 1.92/1,000 catheter-days versus radial 0.25/1,000 catheter-days) (OR, 1.9; 95% confidence interval, 1.15–3.4.). Other studies investigating arterial catheter infection rates have not found a difference between the rate of radial or femoral site infections or of BSIs [50,51,52].
In addition to choosing the best site to decrease infection risk, the best aseptic techniques must be used. Rijnders et al. studied colonization rates of arterial catheter tips and found no difference in the incidence of arterial catheter colonization when the catheter was inserted under maximal sterile barrier precautions, defined as an HCW wearing sterile gloves and a sterile gown, a mask, and a cap with the use of a large sterile sheet; skin was disinfected with 0.5% chlorhexidine in 70% alcohol versus the standard of care group in which hand washing was done, sterile gloves were worn, and the same skin disinfection was applied [46]. However, similar studies of CVC insertions have shown definite infection control value of maximal barrier precautions whose benefit could be operator dependent (e.g., of more benefit with less experienced inserters). Arterial catheters are frequently accessed for blood sampling, so perhaps the focus for decreasing the rate of infection should be on protecting the catheter hub during manipulation [48,53].
The infectious complications of arterial catheters also have been studied in neonates. In different centers, the incidence of colonization of indwelling umbilical artery catheters has varied from 6–60% [54,55,56,57]. Unexpectedly, however, the incidence of colonization fails to increase with duration of catheterization, which suggests that catheters become contaminated initially or soon after insertion through the umbilical stump, an area that is heavily colonized and impossible to sterilize completely by local or systemic antibiotics. Indeed, the same organisms usually are isolated from both the umbilical cord and catheter in any individual patient. The most frequent contaminants are staphylococci, streptococci, and gram-negative bacilli, particularly Pseudomonas, Proteus, Escherichia coli, and Klebsiella. The clinical significance of umbilical catheter colonization is difficult to assess because the incidence of sepsis in most studies has been low. When serial prospective blood cultures have been obtained from umbilical catheterized neonates, however, transient catheter-related BSI has been noted. In a prospective study of temporary (2–4 hours) umbilical catheterization for exchange transfusion, investigators documented a 60% incidence of catheter contamination and a 10% incidence of transient BSI due to Staphylococcus epidermidis (and, in one patient, Proteus) that occurred 4–6 hours after transfusion; this study suggests that the risk of BSI from umbilical catheterization could be highest during catheter insertion and removal [58]. This study and others found that prophylactic systemic antibiotics failed to reduce the incidence of catheter contamination and BSI. At present, systemic antibiotic prophylaxis does not appear to be beneficial during umbilical catheterization; instead, attention should be focused on meticulous cord preparation and care. Other infectious complications of umbilical arterial catheters include mycotic aneurysm or pseudoaneurysm with or without hemoperitoneum [59,60,61].
Pulmonary artery catheters that are inserted through a Teflon® introducer and are mostly heparin bonded have similar rates of infection as do CVCs, but flow-directed
P.782
pulmonary artery catheters carry the added risk of right-sided endocarditis related to endocardial trauma and septic thrombosis of the great vein or pulmonary artery. One autopsy study found that 7% of 55 patients had endocarditis in association with these catheters [62]. Studies that used multivariate analysis have identified a number of risk factors for infection associated with the use of pulmonary artery catheters [63]. Strong independent predictors of an increased risk of catheter colonization were the use of catheters in neonates and in younger children, the placement of the catheter without using maximal barrier precautions, the placement in an internal jugular (rather than a subclavian) vein, the heavy cutaneous colonization at the insertion site, and a prolonged catheterization—particularly >3 days [41,64,65,66,67].
Transducers
Pressure-monitoring devices are used regularly to monitor cardiovascular pressures of critically ill patients. Guidelines for preventing infections related to intravascular pressure monitoring have been formulated and updated by the Centers for Disease Control and Prevention (CDC) [68]. Reusable transducers have been sources of HAIs in outbreaks of gram-negative BSI, candidemia, and dialysis-associated hepatitis [69]. However, disposable pressure transducers can be safely used without change for 4 days, even in busy ICUs [70]. Currently, it is recommended that disposable transducer assemblies be used and replaced every 96 hours [48].
|
Figure 45-1 Heartmate left ventricular assist device (with verbal permission from Thoratec Corporation, Pleasanton, California). |
Circulatory Assist Devices
Left ventricular assist devices (LVAD) are used to maintain a patient's circulation while awaiting cardiac transplantation (see Chapter 38). The frequently used Heartmate (Thoratec Corporation, Pleasanton, California) LVAD consists of a titanium-encased blood pump, an inflow cannula from the left ventricular apex, and an outflow cannula to the ascending aorta. Two porcine valves maintain unidirectional flow. The pump is placed into either a preperitoneal pocket or the abdominal cavity. A driveline connects the device to an external power source via an exit site in the abdominal wall (Figure 45-1). These devices have supplanted the totally artificial heart whose morbidity and mortality are associated with wound infection [70]. Although LVADs may be life saving, infection has been one of its most common complications [71,72,73,74,75]. Infections associated with LVADs include device-related BSI, exit site infections, mediastinitis, and infections at the entrance site of the transcutaneous power cables. One of the more common and difficult problems is a result of the tunneled driveline that connects data between the pump and the extracorporeal controller unit and can result in deep driveline tract, pocket, and device infections. Infectious complications of LVADs are particularly important because they could preclude subsequent heart transplantation [73,74,75,76,77]. In a retrospective analysis of experience with LVADs over an 8-year period, 32% (14/44) of patients did not receive a donor organ because of LVAD-related infection. However, in another retrospective analysis of 14 patients who required an LVAD despite four device-related infections in 3 patients, all 14 patients underwent transplantation that was not delayed because of infection [78]. Others have found the same results [79].
In a particularly robust 9-year retrospective study, investigators found 76 patients who underwent LVAD implantation with the pneumatically or electrically driven Heartmate LVAD as a bridge to cardiac transplantation [73]. Forty-six LVAD-related infections developed in 38 patients for an incidence of 4.9 LVAD-related infections per 1,000 LVAD-days. Twenty-nine LVAD-related BSIs were found in 24 patients for an incidence of 3.1 LVAD-related BSI per 1,000 LVAD-days; persistent sepsis with multiorgan failure caused by LVAD-related infection was the cause of death in 3 patients who did not survive to transplantation. Seven patients who were treated for LVAD-related infection died due to a new infectious complication in the early posttransplantion period versus one death among patients without LVAD-related infection. The driveline exit site appeared to be the major portal of entry in 18/30 episodes of LVAD-related BSI. The most common pathogens were Staphylococcus epidermidis (38%) and S. aureus (24%). Of the LVAD recipients, 84% survived to transplantation. Diabetes mellitus was a risk factor for LVAD-related BSI, and longer device support time was associated with LVAD-related
P.783
infection. Conversely, patients who received continuous antimicrobial therapy from the time of initial diagnosis until transplantation or death had fewer episodes of relapse of infection, of LVAD superinfecions, and of LVAD support times as opposed to patients who received 2–6-week courses of treatment. Another group found 6.2 surgical site infections (SSI) per 1,000 LVAD-days in 36 LVADs implanted in 35 patients [74]. The Heartmate was used for all patients. Three patients developed deep soft-tissue infections (8.3 infections per 100 LVAD implantations) and 6 patients had organ/space infections (16.7 infections per 100 LVAD implantations). Two patients developed mediastinitis; 26% of patients had infections associated with gram-positive bacteria. Of 35 patients, 30 (86%) underwent successful heart transplantation. Three patients had clinical signs of sepsis despite negative cultures. Hemodialysis was the only infection risk factor identified. Finally, a study comparing different LVADs looked at infection rates of the Jarvik 2000 permanent LVAD (Jarvik Heart, Inc., New York) and the Heartmate single-lead vented, electrical LVAD (Thoratec Corporation, Pleasanton, California). The Heartmate group had an infection rate of 0.43 infections/100 patient-days versus 0.08 infections/100 patient-days for the Jarvik group. The Jarvik 2000 LVAD is an axial flow pump designed for permanent use. It has a novel power supply that is a small percutaneous retroauricular skull-mounted pedestal. The investigators believe that the percutaneous, immobile pedestal protected the patients from infection. However, the Jarvik 2000 cannot completely replace left-ventricular function and cannot be used in all patients who need mechanical support [80].
LVADs are associated with a significant risk of infection but still could be superior to medical management alone. Homan et al. prospectively randomized patients to receive LVADs or optimal medical management (OMM) for end-stage heart failure. They found that survival with LVAD (68 patients) was superior to OMM for (61 patients) a 47% decrease death risk (ρ < 0.001) [81].
Uncommon infectious complications of LVADs also have been described. Kotschet et al. described two patients with severe postpartum cardiomyopathy who developed a left ventricular pseudoarneuysm after device removal; the patients had bacteremia during device support [82]. Others have reported fungal infection of the inflow and outflow valves that led to LVAD malfunction and death [83].
Some newer interventions could help to prevent or heal LVAD-related infections. The previously described driveline infections can be difficult to manage. Yuh et al. presented a patient with a deep driveline infection who was successfully treated with a vacuum-assisted closure system [77]. The patient was found to have an abscess within the driveline tract that grew Pseudomonas sp. A vacuum-assisted closure system applied to the wound after debridement was wrapped around the exposed driveline, which was left within the opened tract until adequate healing had occurred. Another possible intervention is an antimicrobial-coated driveline to prevent early infections. Choi et al. evaluated the ability of an LVAD driveline impregnated with chlorhexidine, triclosan, and silver sulfadiazine to resist bacterial and fungal colonization by placing driveline segments onto agar plates inoculated with S. aureus, S. epidermidis, Enterobacter aerogenes, Pseudomonas aeruginosa, and Candida albicans [84]. Antimicrobial activity was demonstrated against all organisms for >14 days, and for >21 days for gram-positive bacteria. In vivo efficacy was tested using rats; 100% of control segments were colonized versus 13% of the test explants. To date, Thoratec has not incorporated this technology [personal communication, October 2006] but has published infection control guidelines for the Heartmate XVE Left Ventricular Assist System. The guidelines were first created by the REMATCH Trial Surgical Working Group and then were updated by the Park City Trial Surgical Working Group in February 2004. They can be accessed at www.Thoratec.com.
Transfusion-Associated Infections
Blood Transfusion and Bacteremia
The three main postulated mechanisms of bacterial contamination of blood products are the use of nonsterile tubing or collection bags due to improper manufacturing, bacteria from the donor's skin or blood, and unsterile handling during preparation and/or storage [85]. Now that systematic blood donor programs have greatly reduced the frequency of transfusion-transmitted viral infections by carefully screening donors and using nucleic acid testing (for HIV and HCV), transfusion-associated bacterial contamination is the most frequent transfusion-transmitted infection. The first case-reports of transfusion-related sepsis appeared in the 1940s and 1950s and involved shock syndromes produced by transfusion of cold-stored blood contaminated with psychrophilic organisms able to survive and grow at 4°C (30°F), such as Achromobacter and some Pseudomonas species. Prospective microbiologic studies soon followed these reports and documented a contamination rate of 1–6% in banked blood [86]. Most contaminants were normal skin flora, presumably introduced with fragments of donor skin cored out during phlebotomy. Such contaminants usually were present in extremely low concentrations (several logarithmic factors below the level of ~100 organisms per milliliter of blood associated with transfusion sepsis), and multiplication of organisms during storage seemed unlikely because of the long lag phase produced by refrigeration and of the antibacterial action of blood. Indeed, retrospective studies failed to document any clinical illness associated with the transfusion of blood that contained low-level skin flora contamination [87]. Nevertheless, asymptomatic patients or patients with nonspecific gastrointestinal symptoms on rare occasions still could be
P.784
a source of bacterial contamination, especially of Yersinia enterocolitica-contaminated red blood cell transfusions. Infections due to this contamination have been associated with a high mortality rate, particularly with units stored >25 days at 1–6°C (34–43°F). The donors presumably had asymptomatic bacteremia at the time of donation. An example of bacterial contamination of blood components during collection or processing is illustrated by an outbreak of Serratia marcescens traced to the use of blood bags intrinsically contaminated during manufacturing [88].
Investigators more recently sought to determine the risk of bacterial contamination of blood components by combining data reported to the CDC from blood collection facilities and transfusion services affiliated with the American Red Cross, American Association of Blood Banks (AABB), and Department of Defense blood programs from 1998–2000. A case was defined as any transfusion reaction meeting clinical criteria in which the same bacterial species was cultured from a blood component and from recipient blood and molecular typing confirmed the organism pair as identical. There were 34 cases and 9 deaths. The rate of transfusion-transmitted bacteremia (in events per million units) was 9.98 for single-donor platelets, 10.64 for pooled platelets, and 0.21 for red blood cells (RBC) units; for fatal reactions, the rates were 1.94, 2.22, and 0.13, respectively. Patients at greatest risk for death received components containing gram-negative organisms (OR, 7.5; 95% CI, 1.3–64.2) [89].
The French BACTHEM study assessed transfusion-associated bacterial contamination determinants using a matched case-control study design. Cases were derived from a database of patients presenting during a 3-year period with a transfusion-related adverse event reported to the French blood agency as a suspected case of transfusion-associated bacterial contamination. Of the 158 cases of suspected transfusion-associated bacterial contamination reported during the study period, 41 cases and 82 matched controls were included. The bacteria were gram negatives (42%), gram-positive cocci (28%), gram-positive rods (21%), and others (9%). The overall incidence rate of contamination was 6.9 per million units issued. The risk of contamination was >12 times higher after platelet pool transfusion and 5.5 times higher after aphaeresis platelet transfusion than after RBC transfusion. Gram-negative rods accounted for nearly 50% of the bacterial species involved and for six deaths. Risk factors included patients receiving RBC for pancytopenia, platelets for thrombocytopenia and pancytopenia, immunosuppressive treatment, shelf life more >1 day for platelets or 8 days for red blood cells, and >20 previous donations by donors [90].
Blood Transfusion and Parasitemia
The frequent use of blood transfusions and the increased travel to-from countries where malaria is endemic have led to an increased occurrence of transfusion-related malaria. During the period 1911–1950, ~350 episodes of transfusion-associated malaria were reported worldwide. In contrast, during the period 1950–1972, the number of reported episodes was >2,000 [91]. In the United States, 101 episodes of transfusion-induced malaria were reported during 1957–1994 [92,93]. The United States still has two–three episodes of transfusion-transmitted malaria per year [94,95].
Based on worldwide incidence data, Plasmodium malariae appears to be the most common cause of transfusion-associated malaria, accounting for almost 50% of episodes.Plasmodium vivax and Plasmodium falciparum are second and third in worldwide incidence, respectively. This ordering probably reflects the fact that although P. malariae infection can persist in an asymptomatic donor for many years, the longevity of P. vivax malaria in humans rarely exceeds 3 years and that of P. falciparum rarely 1 year. Hence, there is higher chance for an asymptomatic donor infected with P. malariae to escape detection and become the source of a contaminated transfusion.
The AABB adopted recommended guidelines for the selection of blood donors to prevent transmission of malaria in 1970, but they were relaxed in 1974 [92,93]. In the changes added to the 24th edition of the Standards for Blood Banks and Transfusion Services (effective November 1, 2006), prospective donors who have a definite history of malaria are deferred for 3 years after becoming asymptomatic. Individuals who have lived for ≥5 consecutive years in areas considered malaria endemic by the CDC are deferred 3 years after departure from the area(s). Because platelet and leukocyte preparations also have been incriminated in the transmission of malaria, the guidelines must be applied to potential donors of any formed elements of blood.
Chagas' disease (American trypanosomiasis) is prevalent through South and Central America. The potential for blood-borne transmission is high because some infected individuals can become asymptomatic but still have persistent parasitemia for 10–30 years. After 10 days of storage the infectivity of blood contaminated with this parasite declines, but storage is not a useful method for preventing transmission; moreover, the parasite is viable in whole blood and RBC stored at 4°C (30°F) for ≥21 days. Serologic screening blood of donors has become mandatory in many South American countries. The problem of transfusion-associated Chagas' disease also has become an issue for U.S. blood banks, secondary to increased immigration and to more potentially infectious U.S. blood donors. It is estimated that about 100,000 infected people live in the United States [96], and estimates of T. cruziseroprevalence in U.S. blood donors range from 0.01–0.20%. Six episodes of transfusion-transmitted Chagas' disease have been reported in the United States since the mid-1980s [97,98]. Anyone with a history of Chagas' disease must be permanently prevented from donating blood [99].
P.785
Toxoplasmosis also can be transmitted via blood transfusion. One prospective survey of thalassemia patients who were frequently transfused detected subclinical toxoplasmosis at a rate comparable to that seen in a control group and therefore was considered to be evidence against the transmission of toxoplasmosis by transfusion [100]. Another study found, however, that patients treated for acute leukemia acquired toxoplasmosis after leukocyte transfusions from donors with chronic myelogenous leukemia; serologic data retrospectively obtained from donors revealed elevated anti-Toxoplasma antibody titers [101]. This inferential evidence for transfusion-associated toxoplasmosis is supported by the findings that the disease can be transmitted between animals and by transfusion, that Toxoplasma organisms retain their viability in stored blood for 50 days, and that organisms can be recovered from the blood buffy-coat layers of patients with toxoplasmosis. Because it seems likely that toxoplasmosis can be transmitted if large concentrations of leukocytes are transfused and all of the leukocyte donors had chronic myelogenous leukemia, it is recommended that blood or leukocytes from patients with leukemia not be used especially because recipients' host defenses usually are severely compromised.
As the rate of human Babesia microti has risen in the United States, so has transfusion-transmitted Babesia. Donations from a group of blood donors in Babesia-endemic areas of Connecticut were seropositive 1.4% of the time, and >50% of those had demonstrable parasitemia [102]. More than 40 U.S. episodes of Babesia microti infection acquired by blood transfusion have been reported, including one in an infant [103,104,105]. However, symptoms secondary to disease with Babesia microti can be mild in immunocompetent people, and the true incidence of the infection is unknown. Donors with a history of babesiosis are indefinitely restricted from donating blood because of the possibility of ongoing asymptomatic parasitemia [99].
Because of the many Americans serving in Iraq, there has been the concern for transmission of Leishmanaiasis. As a result, there is a 1-year donor deferral for military personnel serving in Iraq.
As molecular technology, such as polymerase chain reaction, becomes more sophisticated and available, we could be better equipped to efficiently identify parasites in donated blood and prevent transfusion-associated transmission [106,107].
Platelet Transfusion
Approximately 9 million platelet-unit concentrates are estimated to be transfused in the United States each year and 1 in 1,000–3,000 platelet units is estimated to be contaminated with bacteria, resulting in possible transfusion-associated sepsis [108,109]. In fact, the largest prospective study of transfusion-transmitted bacterial infection in the United States confirms that the incidence of bacterial contamination of platelets ranges from 0.04–1.00 [89]. As noted, screening for viral pathogens in blood has greatly improved, and the risk of platelet-associated transfusion sepsis is about 24fold higher than the transfusion risk for HCV and 28fold higher than the transmission risk for HIV [110]. In fact, transfusion-transmitted bacterial contamination of platelets is the most common cause of fatalities related to transfusion-transmitted disease in developed countries. It is estimated that the risk of bacterial-related death after the transfusion of a platelet unit ranges from 1:7500 to 1:500,000 [111,112].
Because it is now recommended that platelets be stored at room temperature (20–24°C/68–75°F) to increase in vivo half-life, concern over the true incidence of intrinsic contamination and the possible proliferation of contaminants during storage is justified. Of interest platelets, historically have been stored at 4°C (39°F). In 1969, Murphy and Gardner demonstrated that platelet storage at 22°C (72°F) led to improved in vivo viability and function as compared to storage at 13°, 20°, and 37°C (55°, 68°, 99°F) [113]. These observations led to the current practice of storing platelets at room temperature for up to 5 days. It seems reasonable to assume that platelet concentrates are as susceptible to contamination during collection as is blood, which is routinely found to have a 1–6% incidence of low-level contamination. Moreover, platelet concentrates, unlike blood, have no protective antibacterial activity, and platelet transfusions are frequently obtained by pooling the contributions of several donors, which additionally increases the risk of contamination. Despite this seemingly negative picture, most bacterial contaminants isolated from platelet concentrates have been normal skin flora, such as S. epidermidis and diphtheroids, present in extremely low concentrations. Even in the highly susceptible patient populations that normally receive platelet transfusions, such contaminants have failed to produce any documented adverse reactions [114,115,116]. However, there has been at least one report of a Gram-positive organism causing septic shock in a young woman. This was an episode of Streptococcus bovis septicemia found to be secondary to contaminated donor platelets. The donor had undergone colonoscopy 2 months before the platelet transfusion [117].
Although meticulous blood-banking techniques and the widespread use of closed collection systems have made platelet transfusion relatively safe, the occurrence of outbreaks emphasizes the possibility of sporadic, significant contamination of platelets. One outbreak involved seven episodes of Salmonella choleraesuis sepsis traced to platelet transfusions from a blood donor with clinically unapparent Salmonella sp. osteomyelitis and intermittent asymptomatic bacteremia [118]. A long incubation period in this outbreak—that is, a mean interval of 9 days between the transfusion with contaminated platelets and the signs of sepsis—was caused by coincidental administration of antibiotics at the time of platelet transfusion in several patients and delayed recognition of platelets as the vehicle of infection. A second outbreak involved two episodes of
P.786
transfusion-induced Enterobacter cloacae sepsis [114]. An investigation revealed that 20% of the platelet pools prepared in the affected hospital were contaminated. Although most of the contaminants were nonpathogens present only in low concentrations, 6/258 platelet pools grew E. cloacae. The source of these unusual contaminants was not discovered. A third outbreak with Serratia sp. was traced to contaminated evacuated tubes used after blood collection [119]. In a fourth outbreak [120], a cluster of four patients at a university hospital received platelets contaminated with Bacillus cereus, Pseudomonas aeruginosa, or S. epidermidis during a 34-day period. The patient with platelet transfusion-related Pseudomonassepsis died; the remaining patients survived. The investigators surmised that the most likely explanation for the outbreak was contamination of the platelets at the time of phlebotomy. In addition, the four contaminated platelet units were significantly older (mean age, 4.8 days) than 106 randomly selected individual platelet units (mean age, 3.7 days; ρ = 0.04). The hospital increased its surveillance, probably after the one patient died, which could have fostered the discovery of the contaminated platelet pools. More recently, two episodes of Salmonella sp. sepsis, one fatal, from platelet transfusions linked to an asymptomatic bacteremic donor presumably infected during handling his pet boa constrictor was reported, and in 2004 two fatal episodes of transfusion-associated sepsis occurred in platelet recipients [112,121]. One patient, a 74-year-old man with leukemia, had received a transfusion consisting of a pool of five platelet unit concentrates. The pooled platelet unit had been tested for bacterial contamination with a reagent strip test (Multistix®BayerDiagnostics, Tarrytown, New York) before transfusion to determine the pH level, a means for detecting the presence of bacteria. The pH test result was within the accepted range for quality control of the clinic's blood bank. After the transfusion, the patient's blood grew S. aureus, and the patient died 21 days after hospital admission. The same organism was cultured from the leftover platelet unit bag and was indistinguishable by pulsed-field gel electrophoresis (PFGE). The second patient was a 79-year-old man who had received a transfusion of pheresis platelets for thrombocytopenia after coronary artery bypass surgery. The platelets were tested for bacterial contamination with liquid culture media (BacT/Alert®, Bio Merieux Inc., Durham, North Carolina) and found to be negative after 5 days of incubation. Approximately 1 hour after transfusion, the patient deteriorated and died 72 hours later. S. lugdunenis was cultured from the patient's blood and the leftover platelet bag and were identical by PFGE. These last two episodes highlight some of the opportunities for platelet contamination and difficulties with contamination recognition. Blood collection centers culture single-donor platelets, but pooling platelets is done immediately before transfusion; therefore, the hospital is responsible for the bacterial testing of these units. Because of the logistic problem with culturing and providing platelets on a timely manner, some hospitals use nonculture-based methods such as the pH indicators (used in one of the case-patients just described). Non–culture based methods can result in false-negative results as can culture technique.
If the possibility of transfusion-associated bacterial sepsis is considered (e.g., hypotension or fever occur), the transfusion should be stopped immediately, the reaction should be reported to the hospital blood bank, and an investigation should be initiated. The platelet bag and its contents should be returned to the blood bank for inspection, bag defects, and Gram stain and culture of contents. The patient should have blood cultures drawn. The blood supplier also should be notified of possible bacterial contamination in order to recall and culture blood components from the same donation to prevent additional potential morbidity and mortality.
Clearly, we need to use current technology to better prevent and detect bacterial contamination of blood products so that we can improve safety for patients and possibly extend the shelf life of platelets. It is possible that using pathogen inactivation techniques and a spore-based biosensor to detect low levels of bacteria in real time (e.g., the label-free exponential signal-amplification system [LEXAS], which exploits the spore's ability to produce fluorescence when sensing neighboring bacterial cell), could extend shelf life [122,123]. Since March 1, 2004, all platelets have been tested for bacterial contamination before transfusion as required by both the College of American Pathologists and the standards of the AABB [99].
Even when all precautions are taken, transfusion reactions frequently are unrecognized. HCWs must be aware of the risk of bacterial contamination of blood products, particularly platelets because it is often not considered in the differential diagnosis at the time of transfusion reaction due to the similarity of signs and symptoms to those expected from sepsis from other causes [120]. To assess clinician experience with transfusion-associated bacterial infections and knowledge of the new AABB standard, the Infectious Diseases Society of America (IDSA) surveyed U.S. infectious disease consultants via e-mail and fax during July 27–August 24, 2004. The survey went to all 870 infectious disease consultant members of the Emerging Infections Network, a sentinel provider network of ISDA [124]. Completed surveys were received from 46% (399/870) members. Forty-eight (12%) respondents recalled consulting on 85 reactions to blood transfusions potentially caused by bacterial contamination in which 10 reactions were fatal. In 31% (26) of episodes, contamination was confirmed by positive cultures of the recipient's blood and transfused unit. Of respondents, 20% (78) indicated that they were familiar with the new AABB standard for bacterial detection in platelets.
Unfortunately, several interventions to decrease blood contamination have not been successful; it is possible that in one study, expanding the screening questions was not effective in identifying blood donors who may harbor
P.787
To improve bacterial testing and reporting, the AABB provided additional guidance on standardized definitions for test results, investigation and management of implicated units and associated cocomponents, and laboratory testing of detected organisms [126]. Transfusion-related fatalities should be reported to the Food and Drug Administration, Center for Biologic and Evaluation Research (fatalities2@cber.fda.gov, phone: 1-301-827-6220, fax: 1-301-827-6748).
Albumin Infusion
Because of faith in commercial manufacturing practices and the extremely low incidence of reactions to albumin infusion, most physicians consider commercial human serum albumin to be a completely safe product. Years ago, however, a nationwide outbreak of albumin-related P. cepacia sepsis emphasized that any commercial product, particularly any blood component, is susceptible to contamination [127]. In addition to emphasizing the risk of infection associated with the infusion of a nonformed blood component, it is worthwhile to note that the albumin outbreak illustrates several general problems in the detection and evaluation of low-frequency contamination of commercial products. First, HAIs caused by low-frequency contaminants can be difficult to distinguish from endemic problems in any one institution. In the initial reporting hospital, the infusion-related infections became apparent only because of the enormous quantity of albumin used. Second, because commercial products usually are prepared and sterilized in bulk lots, being able to trace the distribution of individual suspect lots is important. Third, sterility of an infusion product cannot be ascertained by visual inspection. Despite P. cepacia concentrations of approximately 100 organisms per milliliter, the contaminated albumin was completely clear. Finally, sampling schemes currently used for product quality control can miss some contaminants when present in low frequency, and endotoxin could escape terminal filtration and be missed by currently used pyrogen tests.
Emerging and Reemerging Organisms Associated with Blood Transfusion
Newer technology and improvement in the blood collection, handling, and transfusion process have created a safe blood supply. Emerging and reemerging pathogens that are not yet easily identified either in the donor, as with asymptomatic disease, or in the product itself, could pose a threat to transfusion recipients [128]. Mumps, which has reemerged particularly in the 18–25-year-old group, is asymptomatic in 20% of people, and 50% of the time the symptoms are nonspecific. Although transfusion-acquired mumps virus has never been observed, viremia is known to occur and therefore has the potential for transfusion transmission of mumps to transfusion recipients from donors with unrecognized infection and asymptomatic viremia. Members of the AABB Transfusion-Transmitted Diseases (TTD) Committee and representatives of the U.S. Food and Drug Administration (FDA) have made available recommendations for the prevention of transfusion-associated mumps [129]. West Nile Virus (WNV) is another pathogen that can cause asymptomatic viremia and has been transmitted via blood transfusions. Since 2003, all blood donated in the United States has been screened via nuclear acid testing (NAT) of donor pools for presence of WNV. The WNV Task Force, which includes representatives from AABB, CDC, FDA, U. S. Deparment of Defense, the American Red Cross, America's Blood Centers, Canadian Blood Services and United Blood Services, initiated an electronic data network in 2006 to enhance identification and tracking. The WNV Biovigilance Network collates data on blood donors with suspected infection in the United States and Canada. Data are collected from blood donor screening by nucleic acid testing. The data from the network demonstrates the magnitude of this potential problem; until September 12, 2006, the network had detected 216 confirmed positive donations and 137 with pending interpretation. Updated data can be found at www.aabb.org. Finally, the recognition of mad cow disease and varient Creutzfeldt-Jacob disease (vCJD) has led to policy preventing blood donation if the prospective donor had a 3-year stay in certain countries including Great Britain from 1980–1996 or has had a blood transfusion in certain European counties from January 1, 1980, to the present. Although no episodes of transfusion-related transmission have been reported because vCJD is most probably due to a prion, a theoretical risk of transmission exists.
Infection Hazards Associated with Anesthesia
Severe bacterial infections have been well documented in association with the use of contaminated equipment for local and spinal anesthesia and of contaminated anesthesia machines for delivery of general anesthesia.
Anesthetic Agents
Propofol (Diprivan, Stuart Pharmaceutical, Wilmington, Delaware) is an oil-based anesthetic agent that is not an antimicrobially preserved product under USP standards and can be stored at 4–22°C (40–72°F) (Diprivan injection, 2005 package insert). An investigation of seven outbreaks of postoperative infection or acute febrile illness revealed an association with the receipt of propofol. The extrinsic
P.788
contamination of propofol due to poor aseptic technique by anesthesia personnel compounded by the ability of the oil-based product to support the growth of contaminants or the use of the same syringe for multiple patients was thought to be the cause of these outbreaks [130]. Infection control practitioners must maintain surveillance for infections due to propofol, which also has been approved for use as a sedative agent in the ICU. In addition, it could be necessary for practitioners to investigate the manner in which such products are handled to ensure that aseptic technique is being used. Products that do not contain preservatives or antimicrobials but have the ability to support microbial growth because of the products' properties should not be used in multidose vials.
Surgical-Specific Anesthesia/Analgesia
Continuous peripheral nerve block (CPNB) is effective for postoperative analgesia after orthopedic surgery. However, this procedure that uses a catheter also has been the cause of infection. Patients scheduled to undergo orthopedic surgery performed with a CPNB were prospectively evaluated in a 1-year multicenter study. Cultures from 28.7% of the catheters were positive, and risk factors for local inflammation or infection included postoperative monitoring in the ICU, catheter duration >48 hours, male gender, and absence of antibiotic prophylaxis [131]. Another study evaluated 211 catheters for CPNB and found that 57% of 208 catheters had positive bacterial colonization after 48 hours. The most frequent organisms were S. epidermidis (71%), Enterococcus (10%), and Klebsiella (4%). Three transitory BSIs likely related to the catheter occurred. After 6 weeks, no septic complications were noted [132]. As with most device-related infections, duration is a risk factor, and gram-positive organisms predominate. HCWs must consider this with all such procedures.
It is difficult to provide anesthesia for a prostate biopsy that requires two or more needle punctures through a highly contaminated rectum. Obek et al. prospectively evaluated 100 patients who underwent transrectal ultrasound guided prostate biopsy [133]. The patients were randomized to receive a periprostatic nerve block or no anesthesia. High fever and hospitalization were more frequent in the nerve block group; bacteriuria in postbiopsy urine cultures was significantly more common in the anesthesia group. Prospective randomized trials will be necessary to determine the optimum antibiotic prophylaxis regimen in patients undergoing biopsy with a periprostatic nerve block.
Endotracheal Intubation
A potential hazard of anesthesia is the occurrence of BSI secondary to the passage of an endotracheal tube. The organisms isolated from the blood usually are α-hemolytic streptococci, both aerobic and anaerobic diphtheroids, and other anaerobic organisms that normally colonize the upper respiratory tract. A higher incidence of BSI after nasotracheal intubation could occur, however, than after the less traumatic orotracheal route [134], but antibiotic prophylaxis is not recommended for either route.
Prolonged nasotracheal (or nasogastric) intubation has been associated with a 2–5% incidence of sinusitis, which may be occult [135,136]. The maxillary and sphenoid sinuses are most commonly involved [135], and frequent pathogens include S. aureus, Enterobacter, P. aeruginosa, Hemophilus, pneumococci, and anaerobes [136]. Sterile and occasionally infected middle ear effusions also are common in patients receiving endotracheal intubation and mechanical ventilation [137]. The CDC guidelines for preventing healthcare-associated pneumonia now recommend using orotracheal rather than nasotracheal tubes in patients who receive mechanically assisted ventilation; using noninvasive ventilation to reduce the need for and duration of endotracheal intubation; changing the breathing circuits of ventilators when they malfunction or are visibly contaminated; and, when feasible, using an endoctracheal tube with a dorsal lumen to allow drainage of respiratory secretions [138].
The laryngeal mask airway (LMA) is a reusable device for maintaining the patency of a patient's airway during general anesthesia; it consists of an inflatable silicone mask and rubber connecting tube that is generally reusable. Because of the concern for the accumulation of proteinaceous material on LMAs and therefore the potential for transmitting organisms including prions even after disinfection, a number of investigators have quantified the amount of protein contamination after sterilization and found protein deposits even after sterilization that increased with device use [139,140,141]. Whether the remaining protein deposits pose a risk to patients is unknown, but they clearly represent a concern for potential transmission of pathogens that could be difficult to recognize as sporadic episodes at different institutions and at different times. As new devices made of new materials come to the market, it is incumbent upon the manufacturer to recommend safe cleaning practices and upon the infection control practitioner to pay close attention to such developments.
Endotracheal intubation also can place the HCW at risk because of exposure to oral and respiratory secretions at the time of intubation. Severe acute respiratory syndrome (SARS) is an example of a transmissible disease that can cause significant morbidity and mortality. In fact, one investigation found a higher risk of developing SARS for physicians and nurses performing endotracheal intubation [142]. The outbreak inspired the creation of infection control guidelines for anesthesiology in Canada. These guidelines address routine precautions for non-SARS patients in the operating room, management of SARS patients in the operating room, and emergency tracheal intubation of SARS patients outside the operating room. The guidelines were developed in
P.789
consultation with anesthesiologists, intensivists, infection control staff, and respiratory therapists and provide a good example of the dynamic state of emerging pathogens and the value of a prompt response by a multidisciplinary team [143].
Infections of the Central Nervous System: Reservoirs and Shunts
Serious infection can complicate the insertion or prolonged use of two very important neurosurgical devices: the Ommaya-type subcutaneous reservoir used for administering intrathecal therapy for fungal or neoplastic meningitis and the ventricular shunt used for decompression of hydrocephalus (see Chapter 34). Infectious complications have been observed frequently after the insertion or chronic use of subcutaneous intraventricular reservoirs [144,145,146]. The use of valved catheters for the treatment of hydrocephalus (i.e., to shunt cerecro-spinal fluid [CSF] from the lateral ventricle of the brain to the superior vena cava, the right atrium, or the peritoneum) also has been complicated by a high incidence of infections. In a number of studies, the overall incidence of shunt infections has ranged from 6–23% [147]. Most of these infections are caused by S. aureus and S. epidermidis and occurred within 2 weeks to 2 months after surgery; this incidence emphasizes the importance of intraoperative and perioperative wound or shunt contamination in the pathogenesis of shunt infection. Shunt infections uncommonly are caused by other organisms such as vancomycin-resistant Enterococcus faecium or Group B streptococcus[148,149]. The equal risk of infection in patients with ventriculoatrial or ventriculoperitoneal shunts suggests that transient BSI is a less likely cause of such infections because ventriculoperitoneal shunts are not exposed to the bloodstream [150].
The antimicrobial treatment of shunt infections that complicate hydrocephalus usually is unsatisfactory unless the shunt is removed [151]. Adherence of slime-producing, coagulase-negative staphylococci to shunt material could be only one reason for failure of antimicrobial therapy [152]. Although <10% of patients has the infection eradicated by systemic antimicrobial therapy alone, those treated with combinations of systemic and intraventricular antibiotics can have 30–90% cure rates [153]. Repeated administration of intraventricular antibiotics has its own complications, however, and when infection is widespread, the treatment of choice appears to be the administration of appropriate systemic antibiotics and the complete removal of the shunt and insertion of a new one at another site [154]. Even after appropriate therapy and presumed successful treatment of a CSF shunt infection, reinfection can occur [155]. Preferably, some time should elapse between the removal of the infected shunt and the insertion of a new one. Despite this discouraging picture, many of the antibiotics previously used to treat shunt infections have been supplanted by newer agents that could prove to be more efficacious. In addition, the epidemiologic characteristics of shunt infections (e.g., perioperative acquisition of organisms) and the narrow spectrum of shunt pathogens suggest that the use of prophylactic antimicrobials and antibiotic-impregnated cerebrospinal fluid shunt catheters at the time of shunt surgery could prove beneficial [156,157]. Two meta-analyses [158,159] evaluating the use of perioperative antimicrobial prophylaxis suggested that the use of prophylactic antibiotics is associated with a significant reduction in subsequent CSF shunt infection. Both studies demonstrated a ~50% reduction in infection risk. Only a few of the studies included in these two meta-analyses reached statistical significance by themselves, which could be due to lack of power at least in part. A third meta-analysis that included randomized or quasi-randomized controlled trials also found the use of systemic antibiotic prophylaxis to be associated with a decrease in shunt infection. In addition, the study evaluated the effectiveness of antibiotic-impregnated catheters to prevent shunt infections and found that too was associated with a decrease in shunt infection [160].
Epidural catheters used for anesthesia or pain management and spinal cord and/or dorsal column stimulators used for pain management also have the potential to cause device-related infections. Few reported episodes of discitis or meningitis are associated with spinal anesthesia. There are, however, some case-reports including one of Streptococcus bovisdiscitis after spinal anesthesia for cesarean delivery, Streptococcus sp. meningitis after epidural anesthesia, and Streptococcus salivarius meningitis after spinal anesthesia [161,162,163]. No specific risk factors were identified. Proper aseptic technique is the mainstay for preventing such infections.
A prospective, randomized, controlled trial to assess the efficacy of a chlorhexidine dressing in reducing the microbial flora at the insertion site of epidural catheters found that the use of the antiseptic at the catheter wound site reduced catheter colonization [164]. The trial authors hypothesized that this could reduce the risk of epidural catheter-related infection. A meta-analysis also supports the use of chlorhexine-impregnated dressing to reduce the risk of epidural catheter bacterial colonization and infection. Analysis of eight randomized controlled clinical trials comparing a single type of chlorhexidine-impregnated dressing with placebo and with povidine-iodined concludes that the chlorhexidine-impregnated dressing reduced the risk of epidural (3.6% vs. 35%; OR 0.07; 95% confidence interval, 0.02–0.31, ρ = 0.0005) exit site bacterial colonization and was associated with a trend toward reduction in central nervous system (CNS) infections. Local cutaneous reactions to chlorhexidine-impregnated dressing were reported in 5.6% of patients in three studies; 96% of the reactions occurred in neonatal patients [165].
P.790
Transient BSI from Nonvascular Procedures
The occurrence of transient BSI associated with relatively noninvasive manipulation of colonized mucosa is well recognized [166]. Such BSIs usually last no longer than 5–15 minutes, at its peak can shower 100 organisms per milliliter of blood (although the peak concentration is almost always much less), and is largely asymptomatic. Hundreds of studies have reported on BSIs after oral treatments alone [167]. This section discusses BSIs after diagnostic gastrointestinal procedures, genitourinary instrumentation, and bronchoscopy; BSIs after endotracheal intubation and invasive vascular procedures were covered earlier in this chapter. Table 45-2 presents a summary of the characteristics of BSIs associated with selected nonvascular procedures.
Gastrointestinal Procedures
BSI has been reported as a sequelae to a variety of gastrointestinal procedures, including sigmoidoscopy, colonoscopy, barium enema, esophagoscopy, biopsy of mucosal masses, injection sclerotherapy of esophageal varices, endoscopic retrograde cholangiopancreatography (ERCP), liver biopsy, esophageal dilatation, and rectal examination. Routine antibiotic prophylaxis is recommended only for high-risk patients before sigmoidoscopy or colonoscopy [168]. Host factors also can influence the incidence and outcome of procedure-related BSI. Rare episodes of symptomatic barium enema septicemia in patients with impaired host defenses (acute leukemia) and in patients with active inflammatory bowel disease have been reported.
Although the role of antibiotic prophylaxis for endoscopy procedures is not always certain, other preventive measures, particularly careful disinfection of endoscopes and good aseptic technique, are of definite importance [169]. Many anecdotal reports and outbreaks have highlight the importance of such measures [170,171,172,173,174,175,176,177,178]. In one report, two episodes of Pseudomonas sp. sepsis in leukemic patients undergoing esophagoscopy with mucosal biopsy were traced to exogenous bacteria introduced during biopsy. Cultures of the esophagoscope and of the endoscopy room revealed widespread contamination with enteric organisms, including P. aeruginosa, and it was shown that routine handling of the instruments ignored aseptic technique [176]. In addition, several series of ERCP-related BSI [170,171,172,174] highlight the multiple sources of contamination in the endoscopy suite, particularly the lens irrigation bottles, and the difficulty in disinfecting the levers and many small-bore channels in these sophisticated instruments even when automatic washers are used. In fact, the automatic endoscope “sterilizers” are at times the source of endoscope contamination, particularly by P. aeruginosa. Significant sporadic problems can be easily overlooked for prolonged periods even with established infection control programs. The poor level of endoscope disinfection in many hospitals occur despite established guidelines and the many reported outbreaks [179,180,181]. Still, there is no recommendation for routinely culturing endoscopes. Interestingly, almost all of the ERCP-related episodes are due to P. aeruginosa, most to one particular serotype, 010, that is either very prevalent or has an undisclosed source related to ERCP procedures. Most of the ERCP patients are at particular risk because of obstructions in the biliary and pancreatic ducts that can trap contaminated injectate. Any patient undergoing ERCP who has suspected ductal obstruction should receive antimicrobial prophylaxis [168]. BSI also has been documented in patients who need esophageal dilation of a stricture and in sclerotheapy of esophageal varicies [182,183,184]. However, endoscopic variceal ligation (EVL) has for the most part supplanted esophageal sclerotherapy; still at least six studies have documented BSI associated with EVL (1–25% of the time) [185,186,187,188,189,190]. Although it is an invasive procedure, percutaneous liver biopsy usually is not associated with infection risk. Ultrasound-guided procedures have a mechanical advantage and can impact positively to prevent infection. A prospective cohort study of 500 patients who underwent an ultrasound-guided liver biopsy identified no infectious complications [191]. Antibiotic prophylaxis with liver biopsy is not warranted at present [168].
|
TABLE 45-2 |
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Nasogastric feeding and enteral nutrition have been associated with BSI and with diarrhea and feeding intolerance, particularly in neonates, due to contamination introduced during collection, preparation, and/or administration of formula or human milk [192,193].
P.791
Urologic and Gynecologic Instrumentation
An association among urethral instrumentation, fever, and BSI has been recognized for many years. In various studies, the incidence of BSIs associated with urologic procedures has been 2–80%, with the greatest BSI risk in patients with preexisting urinary tract infections (UTIs), patients undergoing transurethral resection of the prostate, and patients with prostatitis that is evident on histologic section of biopsy specimens [194]. Similar organisms in 50–67% of patients in whom BSI develops after instrumentation have been recovered from both preinstrumentation urine cultures and postinstrumentation blood cultures. The available evidence suggests that the sources of the other 33–50% of postinstrumentation BSIs include occult prostatitis, the introduction of normal urethral flora, and the contamination of equipment or irrigating fluids before or during instrumentation. Careful evaluation and treatment of genitourinary tract infection should occur before instrumentation [195]; appropriate disinfection of equipment and careful aseptic technique are mandatory. Endometrial biopsy and chorionic villus sampling have been associated rarely with BSI or candidemia [196,197,198].
Pulmonary Procedures
Fever and BSI have been documented in patients after rigid-tube and fiberoptic bronchoscopy.
Procedure-Related BSI Conclusions
Two conclusions can be drawn from the studies of procedure-related BSIs just cited: The equipment used for the procedures should be adequately disinfected/sterilized before every use, and the operator should observe proper aseptic technique. Beyond this, it is apparent that carefully planned, prospective, multicenter studies are needed to assess the incidence and clinical significance of procedure-related transient BSI to determine which hosts are at risk of associated sepsis or infection at distal sites; to determine whether specific risks for certain procedures can be sufficiently defined to justify preventive measures, such as antibiotic prophylaxis; and to determine which prophylactic regimens would be most efficacious. Although such studies cannot ever be conducted, the procedures will continue, and we have tried to note situations in which it seems reasonable to “cover” patients [166]. In this regard, it should be noted that for years dental patients with valvular heart disease have received endocarditis prophylaxis, largely on an empiric basis [199], although the specific regimens [200] and even the mechanisms by which prophylaxis could protect [201,202] have been called into question. The 1997 guidelines of the American Heart Association relaxed recommendations [203].
Additional Procedures Associated with Infections
Interventional Radiology
Percutaneous radiologically guided placement of biopsy needles, catheters, and stents for diagnosis and therapy has become commonplace since the mid-1980s [204]. Infectious complications—primarily BSI, organ perforation, and catheter site infection—vary according to the particular procedure, patient risk factors, and experience of the operator and hospital but are no more frequent than those following more invasive procedures [204]. The use of prophylactic antibiotics for interventional radiographic procedures depends on the situation [205]. In addition, when infection is suspected clinically, therapeutic antibiotics should be administered before a procedure.
Laparoscopic Surgery
Laparoscopic surgery, particularly laparoscopic cholecystectomy, is now one of the most common surgical procedures performed in the United States. Recovery is much faster than after conventional surgery. SSIs have been lower with laparoscopic chelecystectomy [206,207]. Because the biliary tract is normally sterile, antimicrobial prophylaxis in biliary surgery has been recommended only for high-risk patients—defined as those who are >60 years of age or who have had either common duct stones, bile duct obstruction, recent acute cholecystitis, or prior operations on the biliary tract. Until risk stratification data exist for patients undergoing laparoscopic cholecystectomy, antimicrobial prophylaxis standards followed for the same procedure done through a traditional incision can be used [195]. A number of studies have concluded that there is no need for prophylactic antibiotis before laparoscopic surgery in low-risk patients [208,209,210,211,212,213,214].
Cystoscopy
In addition to the risk of BSI associated with cystoscopy, a significant risk of UTI is associated with it. Several remarkably similar outbreaks have been reported in which the use of dilute aqueous quaternary ammonium compounds as cystoscope disinfectants was associated with procedure-related UTIs with Pseudomonas species, particular P. cepacia (seeChapter 30). In these outbreaks, the quaternary ammonium compounds either were ineffective in decontaminating the equipment or were themselves actually harboring viable bacteria while being used as disinfectants [215].
Although the risk of infection associated with the use of dilute aqueous quaternary ammonium compounds has been known at least since the mid-1970s, many hospital personnel persist in using these compounds as antiseptics and disinfectants. Such use has most likely resulted in many outbreaks of hospital-associated UTI and BSIs with occasional outbreaks of hospital-associated respiratory tract
P.792
or SSIs. To help decrease the risk of hospital-associated UTI after cystoscopy, it is important that the equipment be thoroughly cleaned and properly disinfected between uses.
Ureteral Stents
Ureteral stents often are necessary for upper urinary tract drainage but can cause significant patient morbidity including infection. Urologic stents were first developed in 1978 and now include softer biomaterials that are more resistant to encrustation and infections. Chew et al. reviewed the potential use of newer stent materials, coatings, and other innovations, such as the potential for drug-eluting stents [216]. Others have investigated the bacteriology of UTI associated with indwelling J ureteral stents and found that they carry a significant risk of bacteriuria and stent colonization but that the sensitivity of a urine culture could be low, and a negative culture does not rule out a colonized stent. The most common isolates were Escherichia coli, Enterococcus spp., Staphylococcus spp., Pseudomonas, and Candida spp., and stent isolates were more resistant to antibiotics than the organism isolated before stent insertion [217,218]. With new devices and device material, we must find ways to accommodate and enhance surveillance to identify possible new mechanisms of infection and infection transmission.
Bronchoscopy and Endoscopy
An estimated 497,000 bronchoscopy procedures were performed in the United States in 1996 [219]. Several outbreaks have highlighted the problems of pulmonary infection and false-positive culture results because of inadequately cleaned fiberoptic bronchoscopes [220,221]. Especially worrisome is a report of the failure of povidone-iodine to kill Mycobacterium tuberculosis (M. tuberculosis) on bronchoscopes [222]. Preparations of this agent intended for skin degerming often are used inadvisably for decontaminating equipment. This experience has reemphasized the need for higher-level disinfection (e.g., with glutaraldehyde) and/or sterilization of these scopes, especially after use on patients who could have tuberculosis [223]. Even with high-level disinfection, M. tuberculosis has been found in bronchoscopy specimen cultures in 3 patients; specimens of all 3 patients were collected by the same bronchoscope. Only the first patient had clinical evidence of disease with M. tuberculosis. Although the hospital's procedures for disinfection, corresponded with most guidelines, the bronchoscope showed patient debris after disinfection, indicating that the manual cleaning was inadequate and was not approved for reprocessing in the hospital's automated endoscope reprocessor system [224]. Failure to perform leak testing led to failure to discover a hole in the sheath of a bronchoscope, which led to inadequate disinfection and transmission of M. tuberculosis to patients via the bronchoscope resulting in infection and pseudoinfections [225]. Other pseudo-outbreaks related to bronchoscopes that were inadequately disinfected or damaged have been reported [226,227,228]. Others report infection or pseudoinfection due to contaminated bronchoscopes [229]. One outbreak was believed to be a result of a manufacturing defect of the biopsy-port caps, and another was due to incorrect connectors joining the bronchoscope suction channel to the STERIS SYSTEM 1® (STERIS Corp., Mentor, Ohio) processor, obstructing peracetic acid flow through the bronchoscope lumen [230,231]. Infection and pseudoinfection from bronchoscopic procedures reflect a number of problems including ineffective cleaning due to poor technique, damaged equipment, difficult-to-clean accessories, ineffective reprocessing equipment (because of errors and the use of improper connectors, and ineffective disinfectants), use of tap water to rinse the scopes, inappropriate storage (e.g., coiling the scopes), and lack of familiarity with national recommendations for reprocessing. Srinivasan et al. distributed a survey to practicing bronchoscopists regarding infection control issues related to bronchoscopy and specific reprocessing recommendations [232]. Medical directors of bronchoscopy suites or attending bronchoscopists completed 46 surveys. Of the respondents, 65% were not familiar with national reprocessing recommendations, and 39% did not know what reprocessing procedure was used at their own institution. In addition, some parts of the bronchoscopes (e.g., reusable spring-operated suction valves) could require autoclaving if they become heavily contaminated with microbes that are relatively resistant to disinfection such as mycobacteria [233].
Cholangits is a complication associated with ERCP, and outbreaks have been the result of such procedures [234]. Despite negative surveillance cultures of the endoscopes, one outbreak of multidrug-resistant Pseudomonas aeruginosa cholangitis after ERCP was reported [235].
Endoscopic ultrasound (EUS) also has been examined for risk of infection after ERCP; one evaluation of an ambulatory endoscopy center found few infections [236]. Prophylactic antibiotics should be given before EUS of pancreatic cystic lesions [168].
Bronchoscopic and gastrointestinal endoscopic procedure-related infection and pseudoinfection are ongoing problems whose full impact is yet to be defined. Because infection can be associated with an endoscope exposed to microorganisms either from the environment or from a previous patient use, several organizations have published guidelines on infection control for flexible endoscopes [237,238,239]. In spite of the many adverse reports noted and published guidelines, various suboptimal procedures for disinfecting endoscopes are being practiced (Table 45-3) [240,241,242,243]. Even if institutions using endoscopes followed the manufacturer's guidelines for disinfecting scopes, it is becoming more evident that perhaps it is not always possible to clean such devices, most likely because of their complex designs. When “inspection endoscopes” were used to examine the conditions of the
P.793
working and suction channels of 241 flexible gastrointestinal endoscopes at 80 healthcare facilities [244], it was found that 47% (38/80) of facilities had at least one patient-ready endoscope whose suction or biopsy channels were visibly encrusted with debris, and 11% (26/241) of endoscopes had severely scratched channels that provided pockets for debris. Only 5.4% (3/56) of facilities that attempted to dry their endoscopes between procedures were successful. Because high-level disinfectants require clean surfaces, flexible endoscopes must be carefully cleaned of all mucus, blood, and other biologic materials before subjecting them to a high-level disinfectant [239,245]. To further complicate endoscope care, automated machines developed for endoscope reprocessing have been flawed [229,246,247,248,249]. Users should adhere carefully to the manufacturer's protocols but also should be aware of the possibility that colonization of the washer holding tanks is not reversible despite use of the manufacturer's recommended disinfection protocol. Surveillance for endoscope-related infection and pseudoinfection is important, and infection control practitioners must educate their endoscope users (e.g., endoscopy suite personnel and physicians) about problems discussed in this section; the users also must be vigilant to monitor best practice for a complicated cleaning procedure because there are many opportunities for inadequate disinfection. As Weber and Rutala noted, preventing outbreaks from endoscopes requires cleaning to precede disinfection or sterilization, avoiding ineffective or inadequate concentrations of disinfectants, contacting all internal and external surfaces with the disinfectant, and attaching all channel connectors according to the manufacturer's directions when an automated endoscope reprocessor is used. Following disinfection, a sterile water rinse followed by forced-air drying or a tap water rinse followed by forced-air drying and a 70% alcohol rinse must be used to prevent recontamination. The disinfected endoscope must be stored in a manner to prevent recontamination [250].
|
TABLE 45-3 |
||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||
Arthrocentesis and Thoracentesis
Although septic arthritis is caused most commonly by hematogenous spread of organisms, sporadic episodes of Staphylococcal spp. arthritis and, at times, gram-negative bacillary arthritis have followed several days after invasive joint manipulations. During the mid-1960s, CDC investigated a cluster of infection episodes of staphylococcal arthritis that the occurred 1–7 days after outpatient arthrocentesis or intraarticular injection of steroids. Epidemiologic evidence suggested that the physician who had performed these procedures was a disseminator of the epidemic strain, and microbiologic investigation showed that areas of chronic dermatitis on the physician's hands harbored the epidemic organism. A similar cluster of staphylococcal arthritis in which the infections occurred 5–6 days after arthrographic examination of the knee joint and 3–4 days after knee surgery was traced epidemiologically to the surgeon who had performed these procedures who was a nasal carrier of staphylococci. In 1987, 10 episodes of Serratia spp. septic arthritis were
P.794
traced to contaminated benzalkonium chloride antiseptic used in a physician's office [251].
Other diagnostic taps such as thoracentesis [252] also have been associated with HAIs. This emphasizes the fact that all invasive procedures should be performed only under strict aseptic conditions with careful skin antisepsis by an appropriately scrubbed and gloved operator using sterile equipment. Although the relative rarity of centesis-associated infections can be considered testimony that good technique usually is observed in hospitals and clinics, the lack of such infections also could be evidence of the capacity of the local tissue response to limit bacterial invasion in uncompromised hosts [253]. When procedures are performed in patients with compromised host defenses or on tissues that could have diminished ability to limit bacterial invasion (e.g., rheumatoid joints), the risk of procedure-associated infections can be considerable, which emphasizes the need for continued vigilance.
Peritoneal Manipulation
Infectious complications of laparoscopy and amniocentesis are rare, presumably because of careful technique, sterile equipment, local host defense mechanisms, and patients' frequently healthy nature. In fact, high-level disinfection of peritoneoscopes with glutaraldehyde instead of gas sterilization has appeared acceptable. In a retrospective analysis of polymicrobial bacterial ascites in 1,578 abdominal paracenteses, only 1 episode of clinical peritonitis developed, presumably due to entry of the bowel by the paracentesis needle [254].
Artificial Insemination
Artificial insemination has transmitted a variety of infections and mandates careful adherence to protocol for screening candidates [255]. Artificial insemination could have resulted in HIV-1 infection of a woman inseminated with her HIV-infected husband's semen in spite of attempts to remove the virus from semen. The CDC recommends against insemination with semen from HIV-infected men [256]. Although there are no updates to the CDC recommendations, some centers do practice insemination with isolated and virologically tested spermatozoa for couples with an HIV-infected male partner [257,258,259]. However, an episode of HIV-1 transmission through artificial insemination was reported from an infertility clinic in India [260].
Ophthalmologic Examination
Manipulation of the conjunctiva and cornea can occur during tonometry, instillation of eye drops, and manual ophthalmologic examination. Such manipulation can result in conjunctivitis and other eye infections. The infection most commonly transmitted is epidemic keratoconjunctivitis, a highly contagious, frequently iatrogenic disease usually caused by adenovirus type 8 [261]. Transmission of the virus occurs through fomites, such as inadequately disinfected tonometers or contaminated eye droppers and by indirect person-to-person spread from HCW hands. Similar modes of transmission have been implicated in outbreaks of other viral and bacterial eye infections. Although proper care of equipment and conscientious hand hygiene between patient contacts is remarkably effective in halting the transmission of such pathogens, some manufacturer recommendations for disinfection could be inadequate [262], and ongoing community outbreaks could require extrastringent triage and infection control to limit nosocomial spread [263].
References
P.795
P.796
P.797
P.798
P.799