Polymicrobial Infection in Patients with Cancer: An Underappreciated and Underreported Entity
- Reprints or correspondence: Dr. Amar Safdar, Dept. of Infectious Diseases, Infection Control and Employee Health, 1515 Holcombe Blvd., Unit 402, Houston, TX 77030 (asafdar{at}mdanderson.org).
Abstract
Polymicrobial infections account for ∼15% of infections in immunocompromised patients with cancer. However, limited information exists regarding the spectrum and microbiology of these infections, even in severely neutropenic patients. Most studies describe only monomicrobial bloodstream infections in detail, and information regarding polymicrobial infections and nonbacteremic infections is often incomplete or not provided at all. The current lack of well-established definitions for various infections in the immunocompromised host, including pneumonia, neutropenic enterocolitis, and polymicrobial infections, probably plays an important role in the paucity of published information. In this review, we briefly describe the limited information available regarding polymicrobial infections in patients with cancer and address the need for establishing consensus definitions for site-specific polymicrobial infections in neutropenic and nonneutropenic patients. We anticipate that, as factual information regarding such infections becomes available, a more comprehensive understanding of the true scope and impact of these infections will emerge, leading to appropriate modifications in the diagnostic work-up and in the therapeutic approaches used in treating these patients.
Neutropenic patients are susceptible to a wide variety of infections (table 1). The spectrum of infection in such patients undergoes periodic changes that have been well documented and that are influenced by the underlying immunologic deficit(s), the nature and intensity of antineoplastic therapy, the use of antimicrobial prophylaxis and empirical antibiotic therapy, the presence of catheters and other medical devices, and local epidemiological factors [1]. It is important to conduct periodic epidemiologic surveillance studies to document changing trends, because these have a significant impact on antimicrobial prophylaxis and empirical antibiotic therapy [2, 3]. Several study groups and investigators have described current trends in the epidemiology of bloodstream infection in patients with hematologic malignancies and solid tumors [4,5,6–7]. Most of these reports have documented a predominance of gram-positive organisms and a substantial decrease in gram-negative organisms as causes of bacterial infection in this patient population. Unfortunately, most of these reports have focused attention only on bloodstream infections caused by a single organism (monomicrobial bloodstream infections) and have not included detailed information on polymicrobial bloodstream infections or infections at other sites. Consequently, these data provide an incomplete picture, because the overall spectrum of infection in patients with cancer may differ considerably from that associated with monomicrobial bloodstream infection alone [1, 8].
Table 1
Common infections in neutropenic patients.
A guideline, published from a consensus report prepared for the Immunocompromised Host Society for the evaluation of new anti-infective drugs in febrile neutropenic patients, states that “it is important to prospectively define and provide standardized criteria for infectious entities that may develop during the clinical trial” [9, p. 5208]. Two broad categories of infection were defined: (1) bloodstream infection or fungemia involving ⩾1 organism without a definable nonhematogenous site of infection and (2) infection at a specific site (e.g., pneumonia and cellulitis) with or without concomitant bloodstream infection or fungemia. Unfortunately, specific definitions for these infections were not provided, leaving it to each site institution or group to create these definitions for specific trials [10].
There are only a few reports specifically addressing polymicrobial infections, such as bloodstream infection, pneumonia, perirectal infection, neutropenic enterocolitis, and urinary tract infection, despite the fact that these infections are not rare [8, 11, 12]. Most information is derived from therapeutic trials in which only organisms causing bloodstream infection are reported, but even in these reports, the data on polymicrobial infection are usually sketchy and incomplete.
Polymicrobial Bloodstream Infection
Data on polymicrobial bloodstream infection from large antibiotic trials or surveillance studies involving febrile neutropenic patients are summarized in table 2. These data demonstrate that the proportion of bacteremic infections that are polymicrobial ranges from 8% to 32% [3, 13,14,15,16,17,18,19,20–21]. It is difficult to compare data from different institutions, because consistent definitions of what constitutes a polymicrobial bloodstream infection are lacking. For example, in one of the earliest studies of infection in patients with acute leukemia, polymicrobial bloodstream infection was defined as the isolation of <1 organism from blood culture specimens collected from a patient during a 1-week period [22]. In a subsequent study that was conducted by some of the same authors 15 years later, bloodstream infection was considered to be polymicrobial when multiple organisms were cultured during a period of 72 h [11]. These differences highlight the need for developing a standard definition of what constitutes a polymicrobial bloodstream infection.
Table 2
Polymicrobial bloodstream infections in selected large antibiotic trials involving neutropenic patients.
Another significant deficiency in interpreting data on polymicrobial bloodstream infection is that reports documenting the frequency of these infections seldom describe the organisms isolated. In a large study of 507 polymicrobial bloodstream infections reported from our institution from 1972 through 1981, 385 (76%) of the infections involved at least 1 gram-negative bacillus, and 168 (33%) of the infections involved only gram-negative bacilli [11]. Polymicrobial infections caused by gram-positive, anaerobic, or fungal organisms (without the presence of a gram-negative bacillus) were infrequent. The gram-negative bacilli isolated most frequently were Escherichia coli (38% of infections), Klebsiella species (28%), Pseudomonas aeruginosa (20%), and Enterobacter species (13%). The most commonly isolated gram-positive organisms were streptococci (25% of infections), enterococci (16%), coagulase-negative staphylococci (13%), and Staphylococcus aureus (12%). Two organisms were isolated in 363 (72%) of the cases, 3 organisms were isolated in 109 (21%), 4 organism were isolated in 27 (5%), and <4 organisms were isolated in 8 (2%). The overall response rate was 50%, which is substantially lower than the response rate associated with monomicrobial infection (65%–94%). The poorest responses were seen in patients with persistent neutropenia (25%), pneumonia (19%), and bloodstream infection caused by multiple gram-negative organisms (49%).
In a large study conducted by the Surveillance and Control of Pathogens of Epidemiologic Importance project, 2340 patients with underlying malignancy and an episode of nosocomial bloodstream infection were identified [3]. Fourteen percent of these infections were polymicrobial, and only 251 patients (10.7%) were neutropenic. No details regarding the microbiological characteristics of these polymicrobial bloodstream infections were provided.
Several other studies deserve comment. Hung et al. [23] reported a study involving 76 febrile episodes in neutropenic children with cancer. of 27 microbiologically documented infections, 4 (15%) were polymicrobial. of interest, most of these infections involved the urinary tract. Norgaard et al. [24] reported 358 episodes of bloodstream infection in patients with hematological malignancies between 1992 and 2002. Only 110 (31%) of the patients were neutropenic, and polymicrobial infection occurred in 50 cases (14%). The only information provided regarding the infecting organisms was that E. coli were isolated from 30 (60%) and S. aureus were isolated from 7 (14%) of the patients with polymicrobial bloodstream infection.
Pneumonia
A total of 15%–30% of documented infections in febrile neutropenic patients are classified as pneumonia [25]. The majority of these are clinically and/or radiographically documented but do not have microbiological confirmation because of (1) the inability of patients to produce adequate sputum specimens and/or (2) the inability of many patients to tolerate diagnostic procedures, such as bronchoalveolar lavage [2]. These obstacles have severely limited our ability to identify the infecting pathogens. A recent review of 492 patients with cancer who underwent bronchoalveolar lavage at our institution demonstrated that multiple organisms were isolated from these specimens in 74 patients (15%). We excluded patients from whom coagulase-negative staphylococci were isolated (unpublished data). Among the combinations of pathogens were gram-negative bacilli (Enterobacteriaceae, Pseudomonas aeruginosa, Stenotrophomonas maltophilia) in ∼15% of patients, a mold (predominantly Aspergillus species) plus a bacterium in 12%, multiple potentially pathogenic fungi in 22%, and a bacterium plus Candida species in the remainder of the patients. Experience at our institution indicates that Candida species are rarely a cause of pneumonia, even in severely immunosuppressed patients [26]; therefore, in most culture samples from which Candida species were isolated, their presence probably did not represent a cause of the infection.
Neutropenic Enterocolitis
Neutropenic enterocolitis (typhlitis) occurs predominantly in children with acute leukemia but can occur in others who experience prolonged neutropenia [27, 28]. The reported incidence among patients with acute leukemia varies from 0.8% to 26%, and the mortality rate approaches 50% [29,30–31]. Neutropenic enterocolitis is believed to be attributable to chemotherapy-induced mucosal injury that facilitates infection by aerobic and anaerobic enteric organisms [27, 32, 33]. The usual symptoms include fever, abdominal pain, tenderness, and diarrhea [29]. Although it is likely that these infections are polymicrobial, the offending pathogens are seldom identified antemortem except when isolated from blood culture specimens. Aerobic gram-negative bacilli are isolated most often, although Clostridium septicum has been associated with this disease, and polymicrobial bloodstream infection is not uncommon in patients with neutropenic enterocolitis [27, 32, 34].
Perirectal Infection
Perirectal infection occurs in ∼10% of patients with acute leukemia and is increasing in frequency among patients with solid tumors [35]. These infections are predominantly polymicrobial, and frequent pathogens include E. coli, P. aeruginosa, Bacteroides species, Clostridium species, Prevotella species, Peptostreptococcus species, S. aureus, and Enterococcus species. [36, 37]. It is often difficult to identify the true pathogens from fecal contaminants, because few patients form abscesses that can be aspirated. Concomitant bloodstream infections, however, are usually polymicrobial.
Other Polymicrobial Infections
The increasing use of hepatobiliary procedures, such as hepatic artery catheterization for embolization and chemotherapy, and various biliary drainage procedures has been associated with infectious complications [38, 39]. The infections most often encountered include cholangitis, bloodstream infections, liver abscess, gallbladder abscess, and subphrenic abscess. The majority of these infections are polymicrobial, with enteric gram-negative bacilli, Enterococcus species (including vancomycin-resistant enterococci), and Candida species being isolated frequently from abscesses, as well as from blood culture specimens.
Similarly, there has been a substantial increase in the frequency of recurrent, complicated urinary tract infection with the increasing use of stents and nephrostomies, as well as other urinary diversion procedures, and the creation of orthopic neobladders following extensive pelvic surgery [40,41,42,43–44]. The frequency of urinary tract infection in such patients is high (12%–39%), and the infections are often polymicrobial. These infections are frequently recurrent, difficult to treat, and may require administration of prolonged suppressive antibiotic regimens.
Discussion and Recommendations
A review of the current literature and our own extensive experience confirm that polymicrobial infections occur frequently in patients with cancer, both in the presence of neutropenia and in patients who are not neutropenic. These infections are given minimal attention in the large clinical trials of therapeutic regimens for neutropenic patients. For example, many laboratories do not process culture specimens from respiratory secretions, voided or catheterized urine, drainage from intra-abdominal or pelvic collections, or infected surgical sites when multiple organisms are recovered. Most of these culture results are simply reported as “normal site flora” or “contaminated specimens”, without identification of the organisms to the species level or the performance of drug-susceptibility testing. Laboratories may also not process specimens or may reject specimens as inadequate if they do not contain an adequate number of inflammatory cells (neutrophils), an issue which is of great importance in neutropenic patients, because these specimens will invariably not contain the requisite number of neutrophils.
A major obstacle to our understanding of these infections is a lack of definitions of what constitutes a polymicrobial infection. table 3 presents some definitions to help resolve this problem. They are, admittedly, preliminary and will require modifications, but they represent a starting point for further deliberations.
Table 3
Suggested definitions for patients with cancer with polymicrobial infections.
Even with the acceptance of uniform definitions, other problems exit. It is important to understand the pathogenic potential of certain organisms that are part of normal skin, orointestinal, or lower urinary tract flora, such as coagulase-negative staphylococci and Candida species, when they are isolated from sites other than the bloodstream. For example, coagulase-negative staphylococci are often contaminants but have been identified as the cause of pneumonia in neutropenic patients [45]. Despite these difficulties, we believe that it is important to accurately describe the spectrum, frequency, and microbiological characteristics of polymicrobial infections in both neutropenic and nonneutropenic patients, especially because such infections, compared with monomicrobial infections, are associated with greater morbidity and mortality [11]. This will result in a more complete understanding of the nature of infection in patients with cancer and may lead to changes in empirical antibiotic therapy when such infections are suspected, as well as to the more judicious use of targeted gram-positive agents, such as vancomycin, linezolid, and daptomycin. Consequently, we recommend that societies such as the Infectious Diseases Society of America and the International Immunocompromised Host Society develop and publish standardized definitions for the various infections that occur in immunocompromised patients and/or patients with cancer. We also recommend that laboratories—at least those in institutions that deal with substantial numbers of patients with cancer—develop separate guidelines and/or protocols for handling, processing, and reporting microbiological specimens and data from patients with cancer, particularly those with neutropenia. Finally, we encourage clinicians caring for such patients to focus on and report the entire spectrum of infections that they encounter.
Acknowledgments
Financial support. National Institutes of Health (Cancer Center support Grant CA16672).
Potential conflicts of interest. A.S. is a member of the Speakers' Bureaus of Merck, Cubist, and Pfizer. K.V.I.R and G.P.B.: no conflicts.
- Received January 16, 2007.
- Accepted March 24, 2007.
- © 2007 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
-
- Invited Articles
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 8.186
5-Yr impact factor: 7.898
For The Media
For Authors
- Services for authors
- Self-archiving Policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
