Healthcare-Associated Infections in Patients With Neoplastic Diseases

Aditya H. Gaur

Patricia M. Flynn

A patient with a diagnosis of underlying cancer is at particular risk of infection related to a multitude of factors including immune suppression from the underlying disease and its treatment (e.g., chemotherapy, radiation, or stem cell transplantation) as well as breach in continuity of skin and mucosal barriers related to permanent central venous catheters (CVCs) and therapy-associated mucositis. The source of such infections can be endogenous (e.g., gastrointestinal tract flora) or exogenous (e.g., air, water, or fomites), and the setting of acquiring such infections can be at home, in the community, or a healthcare facility. Overall, advances in the field of oncology often test the fine balance between more aggressive therapy, leading to improved survival, and complications, predominantly infections. In this chapter, we review healthcare-associated infections (HAIs), that is, infections that patients acquire in a healthcare setting, in patients with an underlying malignancy. The more general term, “healthcare-associated,” is now increasingly used in place of “nosocomial” (1,2). Especially for the oncology patient population that has repeated “encounters” with the healthcare system in the form of frequent hospitalizations, treatment in day-care facilities, and visits to outpatient clinics, the line that differentiates HAIs from community-acquired infections can become particularly blurry, with the actual source of infection often hard to discern. Early diagnosis, treatment, containment, and prevention of HAIs are of great importance to the management of neoplastic diseases.

PATHOGENESIS OF HEALTHCARE-ASSOCIATED INFECTIONS IN PATIENTS WITH NEOPLASTIC DISEASE

The Host

Influencing the pathogenesis of infections, including HAIs, in the cancer patient are a multitude of vulnerabilities in host defenses. First, in most patients, the defect is an iatrogenic impairment of the immune system such as suppression of B lymphocytes and antibody production, impairment of T lymphocytes impeding cell-mediated responses, or neutropenia due to intensive chemotherapy or irradiation. While the risk of bacterial infections in patients with neutropenia has long been recognized and drives the empiric management of febrile patients with low neutrophil counts, there is increasing recognition of the risk for viral infections and related morbidity in patients who are lymphopenic (3).

Second is the role of compromised natural barriers to infection. A breach in the integrity of the skin and mucous membranes is a frequent portal of entry for microbes composing the resident flora of these sites. This is especially true in the neutropenic host. Factors that affect skin and mucosal integrity include chemotherapy-related gastrointestinal tract mucositis, skin graft-versus-host disease, urinary catheters, and insertion of temporary or permanent indwelling CVCs. In addition, a tumor mass may obstruct a vital organ, impair circulation, or invade adjacent tissue, resulting in an altered regional anatomy that provides a nidus for infection.

Third are differing levels of infection risk based on the nature of cancer treatment. Prednisone, although not a cause of neutropenia, is a potent inhibitor of both humoral and cell-mediated immune responses, especially T-lymphocyte activity. Cyclosporin adversely affects T lymphocytes by decreasing CD4+ lymphocytes and interleukin-2 synthesis. Irradiation and malnutrition cause decreases in T-lymphocyte function. Mucositis in recipients of high-dose cytosine arabinoside is well recognized and is associated with increased infection risk from oral/gastrointestinal pathogens such as viridans streptococcal species. While most chemotherapeutic agents are associated with some degree of neutropenia, some chemotherapeutic agents may also influence infection risk via other mechanisms, including lung fibrosis related to busulfan, lymphopenia with rituximab, and cyclophosphamideinduced hemorrhagic cystitis.

The Pathogens Because of the extensive use of antibiotics and antifungal agents, the normal microbial flora is deranged, and selective antimicrobial pressure creates a microbial milieu that poses an infection risk to the compromised host. While methicillin-resistant Staphylococcus aureus (MRSA) infection was always considered a classic example of an HAI, the past decade has shown a marked increase in community-acquired MRSA (4), making an assessment of every MRSA infection as a suspected HAI no longer accurate. Bacteria such as coagulase-negative staphylococci (CoNS), Corynebacterium species, and
Bacillus cereus (5), often regarded as contaminants in sterile site cultures collected from the immunocompetent host, pose a real infectious threat to the immunocompromised host. These infections are often associated with CVCs or prosthetic joint infections. Finally, with advances in the field of molecular microbiology, the ability to diagnose long-recognized viral pathogens such as respiratory syncytial virus (RSV) and influenza and the more recently identified viruses, such as human metapneumovirus (6), has increased tremendously (7).

DEFINING HEALTHCARE-ASSOCIATED INFECTIONS IN PATIENTS WITH NEOPLASTIC DISEASE

When reviewing the literature pertaining to surveillance of HAIs in patients with cancer, one should be cognizant of the criteria used to define such infections and the denominator used to quantify them, characteristics of patient populations primarily in terms of risk factors such as duration of neutropenia, existing infection control policies including antimicrobial prophylaxis, and types of resources available to diagnose infections. All of these influence the reported HAI rate. In two prospective surveillance studies in adult and pediatric hematology-oncology patients from Bonn, Germany, the researchers noted an overall HAI rate of 11 and 10.8 per 1,000 patient days, respectively, with roughly 75% of the infections occurring in patients who were neutropenic (11,61). To ensure comparability of surveillance data, these researchers recommend that all surveillance studies in the cancer population should include infection rates based on number of patient days at risk, where “at risk” may be defined as the period of neutropenia.

The National Healthcare Safety Network (NHSN) definitions for HAIs are widely used in the United States (1). While these Centers for Disease Control and Prevention (CDC) definitions are designed for surveillance purposes for use in all acute-care settings, including subpopulations such as patients with cancer, their interpretation in an oncology setting can sometimes be challenging. Consider adjudicating a bloodstream infection (BSI) in a patient with a CVC as a primary CVC-associated BSI versus a secondary BSI with oral or lower gastrointestinal tract mucositis as the source of infection. In a patient population with high baseline chemotherapy-related morbidity such as gastrointestinal tract mucositis, the adjudication of BSIs as secondary BSI by applying NHSN definitions of oral or intestinal tract infection can become particularly contentious.

The NHSN definitions recommend two or more blood cultures drawn on separate occasions to be positive to meet the criteria for a laboratory-confirmed BSI for common skin contaminants such as CoNS. With peripheral cultures becoming increasingly uncommon in cancer patients who have CVCs, most clinicians would consider two cultures drawn from two lumens of a double-lumen CVC as two separate cultures, which, if positive for CoNS, warrant considering that episode as a BSI. The latter is not clarified in the NHSN definitions, which leave room for variable interpretation.

The CDC NHSN criteria state that, for an infection to be called an HAI, “there must be no evidence that the infection was present or incubating at the time of admission to the acute-care setting” (1). Because infections have variable incubation periods, determining whether an infection was incubating at the time of admission may be difficult. For this reason, many define an HAI as occurring within 24 hours (8), 48 hours (9), or 72 hours after admission.

In conclusion, the above examples highlight the importance of having a standardized definition for HAIs for oncology patients that takes into account the nuances of this patient population. A consistent way of collecting and reporting the numerator and denominator information when it comes to describing HAI rates in this patient population is key to assessing the effectiveness of interventions to reduce HAIs and facilitate crosscenter comparisons. Of note, even with standardized definitions, there is variable interpretation, as shown by a survey of Australian infectioncontrol professionals, who showed concordance of opinion only 62.5% of the time when adjudicating case scenarios as primary versus secondary BSI based on the NHSN definitions (10).

EPIDEMIOLOGY

Cancer patients present the healthcare epidemiologist with several unique challenges. Foremost is the distinction between an HAI and a community-acquired infection. Cancer patients have a high frequency of interaction with the healthcare setting, with frequent outpatient visits and admissions related to chemotherapy and other noninfectious reasons. Under these circumstances, an HAI may be diagnosed when the patient is not in the hospital and vice versa. In addition, the endogenous microbial flora may change after hospitalization. Especially during prolonged hospital courses, microorganisms of the hospital environment may be acquired that will increase the patient’s risk for an infectious episode. Furthermore, because of the multidisciplinary management, some cancer patients may move through many sites during one hospitalization, such as the operating room, intensive care unit, medical service, physical rehabilitation units, and diagnostic imaging and irradiation departments. Under such circumstances, tracking the source of infection sometimes requires exhaustive epidemiologic investigation.

The ability to diagnose viral infections with increased sensitivity using polymerase chain reaction (PCR)-based techniques may complicate distinguishing between reactivation of latent viral infections and new infections. Latent infections acquired early in life may become activated during immunosuppression and hospitalization. These must be differentiated from acute primary infections caused by the same microorganism that could have been acquired during hospitalization. Notable among these are the herpes virus infections, including herpes simplex and cytomegalovirus (CMV) disease. Also, recurrent Varicella zoster virus (VZV) infection in the form of disseminated zoster is sometimes difficult to differentiate from primary varicella. Evidence suggests that some cases of Pneumocystis jiroveci pneumonitis may be acutely acquired infections in the hospital rather than the more usual reactivation of a latent infection.

There are also challenges in establishing the etiology of an infectious episode in the immunocompromised cancer patient. Since most infections are due to commensal or opportunistic microorganisms of the normal microbial flora, the isolation of a microorganism by culture may not necessarily prove it to be the cause of the illness. For example, in the febrile neutropenic patient, a microorganism such as Corynebacterium species isolated from a blood culture may be the causative agent or may merely be a skin commensal or contaminant of the culture. Additionally, the recognition of an infected site can be challenging. In the severely neutropenic and anemic patient with cancer, the key signs of infection may be absent because of a lack of inflammatory response. S. aureus may be introduced in a healthcare setting at the time of a finger stick for a blood count. Without neutrophils, no infiltration occurs, so swelling of the affected finger may be absent; furthermore, the anemia does not allow the appearance of erythema, so the sole manifestation of the infected finger stick site may be local pain or fever. Meningeal infection in the neutropenic patient may lack the typical signs of meningeal inflammation such as a stiff neck and a paucity of neutrophils in the spinal fluid. Even with fairly extensive bacterial infection in the lung parenchyma, the neutropenic patient may not be able to mount a sufficient inflammatory response to create an infiltrate recognizable on a chest radiograph. Finally, as previously discussed, consideration should be given to creating oncology population-specific definitions of HAIs that allow for comparisons within and between institutions.

Thus, the healthcare epidemiologist must consider these and other nuances of the compromised host with cancer when tracking HAIs. Molecular techniques to characterize microbes by subcellular and genetic components are evolving as powerful tools for healthcare epidemiology. Analysis of chromosomal DNA by pulsed field gel electrophoresis, ribotyping, and random primer PCR methods permits more precise characterization than more conventional phenotyping techniques.

TABLE 57-1 Distribution and Incidence Densities of 263 Healthcare-Associated Infections

*HAI*	*N (Proportion)*	ID
All HAIs	263 (100%)	4.80
Bloodstream infections (BSI)	153 (58%)	2.79
Laboratory-confirmed (blood-culture-positive) BSI	138 (52%)	2.52
Blood-culture-negative BSI	15 (6%)	0.27
Radiologically confirmed pneumonia	20 (8%)	0.36
Invasive aspergillosis	26 (10%)	0.47
Respiratory syncytial virus infection	2 (1%)	0.04
Surgical site infection	15 (6%)	0.27
C. difficile-associated enterocolitis	24 (9%)	0.44
Rotavirus-associated enterocolitis	6 (2%)	0.11
Urinary tract infection	8 (3%)	0.15
Ventriculitis related to external CSF drainage	1 (0%)	0.02
Local infections at the central venous access device exit site	8 (3%)	0.15
HAI indicates healthcare-associated infection; N indicates the absolute number; ID indicates incidence density per 1,000 inpatient days.
(Adapted from Simon A, Ammann RA, Bode U, et al. Healthcare-associated infections in pediatric cancer patients: results of a prospective surveillance study from university hospitals in Germany and Switzerland.

ETIOLOGIES OF INFECTION

HAIs in cancer patients can be caused by a variety of infectious microorganisms, but the most common pathogens that have been reported are bacterial, followed by fungal and then viral. This may change, as viral diagnostics have greatly improved and the routine use of molecular amplification techniques for diagnosis of respiratory infections is increasingly mainstream. These improved viral diagnostics are likely to influence diagnosis of polymicrobial infections (viral and bacterial coinfections) and lead to some reduction in the number of episodes categorized as healthcareassociated fever of unknown origin (FUO) (11). In previous reports from oncology centers, bacterial microorganisms were isolated in more than 75% of HAIs, fungal pathogens in approximately 3% to 10%, and viruses in only 2% (12,13). The distribution of 263 HAIs, prospectively assessed across 7 pediatric oncology centers in Switzerland and Germany between 2001 and 2005, is shown in Table 57-1 (8). Of all HAIs in this study, 58% were BSIs. The rate of fungal infections varies between institutions and even among units within an institution. In one oncology intensive care unit, fungal infections accounted for 22% of all their HAIs (14). Another study of HAIs in neutropenic patients reported a rate of 19% (15). Polymicrobial infections are not uncommon in this patient population. Robinson et al. (12) noted multiple isolates
in one-third of their infections. Both multiple bacterial isolates and mixed infections can occur. As mentioned earlier, when making comparisons between studies and centers, one has to keep in mind the differences in definitions, patient populations, and institute characteristics.

Bacterial Infections

The most important bacterial healthcare-associated pathogens are CoNS, S. aureus, Escherichia coli, and Pseudomonas aeruginosa (12,13) (see Chapters 28, 30, 34, and 35). Together, these four microorganisms account for more than half of healthcare-associated bacterial infections in cancer patients.

Gram-Positive Microorganisms S. aureus was the most frequent bacterial isolate in two surveys of healthcareassociated pathogens in cancer patients, accounting for 14% to 18% of isolates (12,13). Surgical sites were most often involved. CoNS infections have increased dramatically over the past decade; these microorganisms are the most common microorganisms isolated from BSIs in some centers (16,17). The rise of these fairly nonpathogenic bacteria has been linked to the use of tunneled CVCs, such as the Hickman catheter.

Viridans streptococci are normal inhabitants of the oropharynx that invade through damaged mucous membranes and cause bacteremia and pneumonia in cancer patients. A syndrome of severe shock and adult respiratory distress syndrome can result. There is a potential causal relationship with cytosine arabinoside administration (18,19).

Clusters of Corynebacterium jeikeium bacteremia have been reported from several cancer centers (20, 21 and 22). Risk factors include immunosuppression and use of plastic devices such as intravenous catheters. Some evidence suggests that patient-to-patient transmission does not occur (22). The microorganism is resistant to multiple antibiotics, and vancomycin is the suggested therapy.

Gram-Negative Microorganisms As a family, Enterobacteriaceae are common pathogens for HAIs in cancer patients. E. coli and Klebsiella pneumoniae predominate (12,23). These microorganisms, along with Serratia species (24), Enterobacter species (25), and Citrobacter species (26), have been isolated in sporadic infections and in epidemics. They are common causes of bacteremia, pneumonia, and urinary tract infections (UTIs). Frequently, patients are already receiving antibiotic therapy when these infections develop (23, 24, 25 and 26). P. aeruginosa is the most notorious pathogen in patients with malignancies. It is associated with healthcare-associated bacteremia, pneumonia, UTIs, and wound infections. Although a frequent healthcare-associated pathogen, it has a special predilection for granulocytopenic hosts. In a review of P. aeruginosa infections in cancer patients in the 1990s, Maschmeyer et al. (27) noted that the proportion of these infections among cases of gram-negative bacteremia over the past two decades has not generally declined, but there were marked local and regional differences in the incidence of infections. Infections with P. aeruginosa account for approximately 10% of all HAIs in cancer patients (12,13,28). In the hospital environment, P. aeruginosa is associated with respiratory equipment, sinks, and fresh fruit and vegetables. Colonization often precedes infection (28,29). Historically, the case fatality rate for P. aeruginosa infections was reported to be as high as 65% to 70%, which was significantly higher than the rate for other gram-negative bacterial infections (29,30). Newer antimicrobial agents with improved anti-Pseudomonas activity have lowered fatality rates (31).

A variety of other gram-negative microorganisms have also been linked with HAIs in cancer patients. The Legionella species are fastidious gram-negative bacilli. Approximately 42% of cancer patients with Legionnaire’s disease are infected in a hospital setting. The use of steroids and neutropenia appears to have causal roles (32). Stenotrophomonas maltophilia (previously Xanthomonas maltophilia) has been reported as a cause of bacteremia, UTI, pneumonia, and wound infections in cancer patients. It is most often detected in patients who have received antibiotics and respiratory therapy. The microorganism has been isolated from hospital sinks and respirators. The association between the use of respiratory equipment and isolation of S. maltophilia from sputum suggests that the equipment may be a significant reservoir for the microorganism (33).

Anaerobes Anaerobes are infrequent healthcare-associated pathogens in the oncology patient and are isolated in <5% of infections. Usually, obvious disruption of normal gastrointestinal barriers is apparent when infections do occur (34).

Antibiotic-Resistant Bacteria Widespread use of antibiotics, both prophylactic and empiric, has resulted in HAIs caused by multiply resistant microorganisms. MRSA, vancomycin-resistant enterococci (35), and fluoroquinoloneresistant enteric microorganisms have been reported to cause significant problems in an oncology population (36, 37 and 38). A single-center retrospective study in cancer patients shows recent receipt of carbapenem therapy as an independent risk factor for vancomycin-resistant Enterococcus faecium bacteremia, and recent receipt of aminoglycoside therapy as an independent risk factor for vancomycin-resistant Enterococcus faecalis bacteremia (39). Prudent use of antibiotics and careful surveillance of this population are necessary to detect and control the spread of these pathogens.

Fungal Infections

Perhaps the most serious infectious threat to the cancer patient is that caused by the opportunistic fungi, especially candidiasis and aspergillosis. The secular trends in the epidemiology of healthcare-associated fungal infections in the United States from 1980 to 1990 have been described (40). During this decade, the National Healthcare-associated Infections Surveillance system hospitals reported 30,477 healthcare-associated fungal infections. During this time, the fungal infection rate increased from 2.0 to 3.8 infections per 1,000 patients discharged. The medical specialty with a high infection rate was oncology, with rates that varied from 8.9 to 10.6 infections per 1,000 discharges. Candida albicans was the most frequently isolated fungal pathogen (59.7%), followed by other Candida species (18.6%).

While C. albicans is the most common fungal pathogen in cancer patients (see Chapter 40), studies have noted
increases in the frequency of other Candida species, including Candida tropicalis, Candida parapsilosis, and Candida krusei (41). Within individual cancer centers, a significant species shift has been noted even within the non-C. albicans group, such as an increase in C. parapsilosis and a decrease in C. tropicalis (42). Overall, these differences between institutions to some extent are influenced by institutional antifungal prophylaxis guidelines, the use of indwelling catheters, and the types of malignancies treated. A study of candidemia in cancer patients from November 1992 to October 1994 found that, of 249 episodes of candidemia, non-albicans candidemia accounted for 64% (101/159) of episodes in patients with hematologic malignancies and 30% (27/90) of the episodes in patients with solid tumors (43).

Fungemia, pneumonia, UTI, or disseminated disease with involvement of the abdominal viscera may occur. Infections are usually preceded by colonization of the gastrointestinal tract with the offending microorganism, but common source outbreaks have also been reported. Risk factors include the use of antibiotics, colonization with the microorganism, neutropenia, and the presence of tunneled CVCs.

While it is clear that the incidence of invasive aspergillosis has been increasing in patients with cancer, especially those with hematologic malignancies and bone marrow transplant recipients (44), controversy exists regarding the definition of healthcare-associated versus communityacquired infection. This is in part due to factors such as an unknown incubation period and size of “infectious” inoculum as well as lack of uniform, reliable methods for environmental sampling in studies that attempt to trace the source of infection (45). The overall case fatality rate of this disease is very high, with the highest being in bone marrow transplant recipients (46). Sites most often involved include the lungs and the paranasal sinuses. Inhalation of conidia (spores) is requisite to the development of this infection. Direct inoculation of Aspergillus species spores from occlusive materials, such as tape, has also been reported.

Although Aspergillus causes a much lower rate of infection than candidiasis, it is the mycosis that has been most convincingly associated with the hospital environment. Outbreaks of healthcare-associated aspergillosis have been reported to be due to hospital construction and renovation activities (47, 48, 49 and 50). Bone marrow transplant patients are especially susceptible. The source of infection is airborne conidia of Aspergillus species often associated with contaminated air-handling systems. Evidence suggesting the hospital water distribution system as an additional indoor source for pathogenic airborne fungi has also been reported (51).

Historically, while C. albicans accounts for the majority of infections in compromised patients, recent epidemiologic trends indicate a shift toward infections by Aspergillus species, non-albicans Candida species, and previously uncommon hyaline filamentous fungi (such as Fusarium species, Acremonium species, and Pseudallescheria boydi), dematiaceous filamentous fungi (such as Bipolaris species and Alternaria species), and yeastlike pathogens (such as Trichosporon species and Malassezia species) (52). These emerging pathogens are increasingly encountered causing life-threatening invasive infections that are often refractory to conventional therapies. Increasing use of antifungal prophylaxis may be linked to the emergence of these microorganisms as well.

Viral Infections

Overall, viruses account for relatively few HAIs. This number is likely to increase as viral diagnostic technology improves. Known HAI pathogens include VZV virus (53,54), RSV (55), influenza, and rotavirus (56). Hepatitis B and hepatitis C have also been reported from other countries as healthcare-associated pathogens in children with cancer (57,58).

CLINICAL MANIFESTATIONS

Fever is the most frequent manifestation of an infection including an HAI in the cancer patient. When fever occurs, especially in the setting of neutropenia, a diagnostic workup, including careful history and physical examination and bacterial and fungal cultures of blood, and any obvious sites of infection such as wounds, should be done before beginning therapy. BSIs most often present with fever with or without evidence of shock. Catheter-related bacteremias or fungemias may present with chills or rigors after flushing the catheter. If a tunneled CVC is in place, all lumens and ports should be cultured. In addition, if symptoms are present, a chest radiograph, and possibly sinus radiographs, should also be obtained. If no source of infection is identified by the diagnostic workup, a diagnosis of FUO may be made.

As noted earlier because of lack of inflammatory response in the neutropenic patient, signs and symptoms may be subtle, and pain and fever may be the only clinical manifestations of a serious HAI. Differentiating an infection from side effects of chemotherapy or radiation can sometimes be very difficult, and, not uncommonly, empiric treatment for an infection is started while waiting for further information.

SITES OF HEALTHCARE-ASSOCIATED INFECTIONS

Bloodstream Infections

BSIs account for a major proportion of HAIs and are associated with an extended hospital stay, extra costs, and excess mortality (59). The mean central line-associated BSI rate between 2006 and 2008 based on data collected by the NHSN from various participating oncology units ranged from 1.7 to 3.9 per 1,000 catheter days for permanent central lines and 2.0 to 4.6 for temporary central lines (60) (Table 57-2). In three prospective surveillance studies of HAIs in adult and pediatric hematology/oncology patients, 43% and 58% of HAIs were BSIs (8,11,61). From March 1995 through February 2001, a total of 22,631 cases of BSI were reported by 49 US hospitals participating in the Surveillance and Control of Pathogens of Epidemiologic Importance (SCOPE) Project (9). Among these cases, 2,711 isolates from 2,340 clinically significant episodes of BSI that met the surveillance definition of HAI were identified in adult patients with malignancies. Of all the recorded
episodes, 61% were caused by gram-positive aerobic microorganisms, and 27% were caused by gram-negative aerobic microorganisms. The proportion of gram-positive microorganisms was 62% for BSIs in 1995 and 76% for those in 2000 (p < .001), indicating a shift from gram-negative to grampositive infections that has been noted over the past two decades by other studies as well. The proportions of gramnegative pathogens during the same periods were 22% and 15%, respectively. The increasing use of tunneled CVCs and the concomitant increase in the number of CoNS infections are some of the factors believed to contribute to this shift toward gram-positive BSIs (62). Another speculation to explain this shift is the widespread use of second-and third-generation cephalosporins for the empiric treatment of febrile neutropenia. Because these antibiotics have improved gram-negative coverage at the expense of grampositive coverage, breakthrough healthcare-associated bacteremias are likely to be of gram-positive origin (63).

TABLE 57-2 Distribution of Laboratory-Confirmed Permanent and Temporary Central Line-associated Bloodstream Healthcare-Associated Infection Rates by Type of Location (NHSN Data 2006 through 2008)

*Permanent Central Line-associated Bloodstream Infection (PCLABSI) rate^a*
*Type of Location*	*No. of Locations^b*	*No. of PCLABSI*	*Permanent Central Line Days*	*Pooled Mean*
Specialty care areas
Bone marrow transplant	21	235	60,546	3.9
Hematology/oncology	41	158	95,535	1.7
Long-term acute care	43 (33)	38	23,278	1.6
Pediatric hematology/oncology	7	75	32,255	2.3
Solid organ transplant	9	11	3,953	2.8
*Temporary Central Line-associated Bloodstream Infection (TCLABSI) rate*^c
*Type of Location*	*No. of locations^b*	*No. of TCLABSI*	*Temporary Central Line Days*	*Pooled Mean*
Specialty care areas
Bone marrow transplant	18 (17)	96	27,290	3.5
Hematology/oncology	33 (31)	117	51,950	2.3
Long-term acute care	67 (64)	260	149,298	1.7
Pediatric hematology/oncology	5	47	10,287	4.6
Solid organ transplant	12	66	32,591	2.0
^a ^bNumber of locations meeting minimum requirements for percentile distributions if less than the total number of locations. If this number was <20, then percentile distributions were not calculated. ^c
(Adapted from Edwards JR, Peterson KD, Mu Y, et al. National Healthcare Safety Network (NHSN) report: data summary for 2006 through 2008, issued December 2009. Am J Infect Control 2009;37(10):783-805.)