Current Spectrum of Bacterial Infections in Patients with Cancer

This is the authors' abstract. We don't add key points for this paper.

Sir—We read with great interest the data presented by Wisplinghoff et al. [1] on current trends in the epidemiology of nosocomial bloodstream infections in patients with hematologic malignancies and solid tumors. Such data provide current information, including information on epidemiologic shifts and the emergence of unusual and/or resistant organisms, and are often used by practicing clinicians to determine the most appropriate agents for antimicrobial prophylaxis and empiric antibiotic therapy in these high-risk patients. We would therefore like to point out that focusing only on nosocomial bloodstream infections (NBIs) and on monomicrobial infections paints an incomplete picture, because the overall spectrum of infection seen in patients with cancer may differ considerably from that associated with NBI. Our institution, a National Cancer Institute–designated Comprehensive Cancer Center, is dedicated exclusively to the treatment of patients with cancer. Our experience over several decades demonstrates that NBIs are neither the most common nor the most lethal infections found among patients with hematologic malignancies or solid tumors [2]. A recent survey at our institution showed that NBI accounted for 35% of infections among patients with hematologic malignancies and 20% of infections among patients with solid tumors (table 1) [3]. Pneumonia was more common than NBIs in both sets of patients, and other common infections included urinary tract, skin and soft-tissue, and gastrointestinal infections (e.g., enterocolitis [typhlitis] and perirectal infections). Although gram-positive organisms caused NBI more often than gram-negative organisms did, there was a predominance of gram-negative organisms at most other sites of infection. Another critical piece of information missing from the current Surveillance and Control of Pathogens of Epidemiologic Importance (SCOPE) study [1] and from surveys conducted by the European Organization for Research and Treatment of Cancer is the proportion of infections among patients with cancer that are polymicrobial [4]. Our data indicate that polymicrobial infections have more than doubled in frequency since the early 1970s and currently account for 23%–31% of documented bacterial infections (table 2) [2, 3]. These are predominantly tissue-based infections (<10% of NBIs are polymicrobial) and are associated with greater morbidity and mortality than are monomicrobial infections [5]. It is more important that ∼80% of polymicrobial infections have a gram-negative component, and ∼33% are caused exclusively by multiple species of gram-negative bacilli [5]. When all sites of infection are taken into account and monomicrobial and polymicrobial infections are included in the overall spectrum, a substantially different epidemiological picture emerges, compared with that associated with NBI alone. In particular, the proportion of gram-positive microbial infections falls dramatically, from the 70%–75% range to <50%. This can have a significant impact on the choice of antimicrobial prophylaxis and on empiric therapy in patients with cancer. Common sites of infection in patients with cancer who have hematologic malignancies and solid tumors. Current distribution of monomicrobial and polymicrobial infections in patients with hematological malignancies and solid tumors. Surveys reporting increasing frequencies of isolation of gram-positive organisms from patients with cancer have been responsible for the drastic increase in the empiric use of vancomycin and, to a lesser extent, of linezolid [6,7]. This has, in turn, led to the emergence of resistant organisms, such as vancomycin-resistant enterococci, vancomycin-intermediate Staphylococcus aureus, and linezolid-resistant isolates, particularly among patients who receive multiple and/or prolonged courses of these agents. Although we commend the SCOPE investigators for their excellent and timely review of the epidemiology of NBI among patients with cancer, we urge them and other investigators conducting similar surveys to focus on the entire spectrum of infections found among such patients. This is likely to result in the accumulation of more complete and accurate data, and, hopefully, in more appropriate antimicrobial usage.