Is Activity against “Atypical” Pathogens Necessary in the Treatment Protocols for Community-Acquired Pneumonia? Issues with Combination Therapy
- Reprints or correspondence: Dr. John G. Bartlett, Johns Hopkins University, School of Medicine, Baltimore, MD 21205 (jb{at}jhmi.edu).
Abstract
The “atypical pathogens” reviewed include Legionella pneumophila, Chlamydophilia pneumoniae, and Mycoplasma pneumoniae. Urinary antigen tests are the most frequently used tests for Legionella species and show good specificity and reasonable sensitivity. For M. pneumoniae, detection of immunoglobulin M, used for the past decade, has substantially improved diagnostic specificity and has simplified testing. For C. pneumoniae, there is no consensus on a simplified test that can be commonly used, and the reported results, with the use of tests that have not been well validated for diagnostic accuracy, show great variation in prevalence. With regard to therapeutic trials, 3 meta-analyses have recently addressed the issue of clinical outcome with or without antibiotics directed against atypical pathogens (macrolides or fluoroquinolones vs. β-lactam agents). These analyses have not been able to demonstrate any clinical benefit, except in a subset analysis of infections caused by L. pneumophila. Nevertheless, multiple studies from the 1950s and 1960s supported a clinical benefit of tetracycline or erythromycin treatment for infections caused by M. pneumoniae. The largest uncontrolled review of antibiotic treatment for hospitalized patients with community-acquired pneumonia demonstrated a clear benefit from the use of macrolides plus cephalosporins or fluoroquinolones, compared with the use of β-lactams alone, although these data support a potential role for atypical agents, because other potential explanations make drawing conclusions difficult. With regard to future studies, it is noted that the standard of care in the United States, Canada, and some other countries is routine use of agents to treat infection with atypical pathogens, which makes the conduct of controlled trials to address these issues ethically difficult and practically impossible. Additional limitations are the difficulty in diagnostic testing for C. pneumoniae and the importance of rapid institution of therapy for patients severely ill enough to require hospitalization. These observations introduce substantial ethical and logistical barriers to studies of specific agents, except by retrospective analyses.
The “atypical pathogens” in community-acquired pneumonia (CAP) that are included in this review are Mycoplasma pneumoniae, Chlamydophilia pneumoniae, and Legionella species. Studies aimed at identifying the etiology of CAP indicate that these atypical pathogens collectively account for ∼20%–30% of all cases of CAP among adults. On the basis of in vitro sensitivity testing and animal models, all 3 of these pathogens appear to have sensitivity to macrolides, tetracyclines, ketolides, and fluoroquinolones [1–10]. β-Lactams are not active against these pathogens. The questions posed in this review concern the need to direct therapy against atypical pathogens in the routine management of CAP. The questions are as follows:
1. Can we diagnose pneumonia caused by specific atypical pathogens?
2. Is there evidence that empirical treatment directed against these pathogens has a sufficiently compelling effect to make it a standard of care?
3. Can therapeutic trials be done to address the issue of the need to treat atypical pathogens in all cases of CAP?
Diagnostic Testing for Atypical Pathogens
The reference standard for identification of infections caused by these 3 atypical pathogens is culture. The contemporary concept is that asymptomatic carriage of these pathogens is rare, so detection of any of the 3 is tantamount to a diagnosis of infection. The issue is the diagnostic accuracy of testing, because of the diversity of and variations in the quality of the tests used. The quality is generally good for Legionella species, is variable for M. pneumoniae, and is poor for C. pneumoniae.
M. pneumoniae. This organism is very difficult to identify by culture [11], so detection requires the use of alternative tests. In the 1960s and 1970s, the most common tests performed in clinical laboratories were serological tests using complement fixation and agglutinins, which were relatively insensitive and nonspecific. Subsequently, there was the use of immunofluorescence assay and ELISA to detect IgM, which eliminated the need for paired specimens [12]. A limitation of the immunofluorescence assay was the need to separate IgG, and the concern about ELISA was the complexity of the testing [13]. The test that has subsequently gained favor is an EIA that detects IgM antibodies to M. pneumoniae, does not require separation from IgG, is rapid (duration, 10 min), and shows results that are comparable to those of previously used complicated tests in children. Overall results for 694 specimens with multiple assays showed that the Mycoplasma IgM assay had 90% sensitivity, 93% specificity, and 92% agreement with “consensus results” [13].
Legionella. The genus Legionella includes >50 known species and at least 16 serogroups of Legionella pneumophila. However, 80%–90% of culture-positive specimens from patients with CAP implicate L. pneumophila [14–16]. Detection methods include culture, serological analysis, urinary antigen tests, and PCR [15]. The yield of culture varies from 15% to >90%, depending, to a large extent, on the severity of the pneumonia [17]. The advantages of culture are the 100% specificity and the ability to detect all these species and serotypes, but the problems are that at least half of the patients in many studies do not produce sputum samples; that most of the specimens would be discarded because of a lack of purulence screens, which is standard practice in many laboratories [18]; and, most importantly, that results are delayed for several days, with some species requiring ⩾10 days of incubation. Consequently, culture is good for epidemiological review but is unrealistic for clinical guidance [15, 17, 19]. Serological analysis shows a sensitivity of 40%–90% [20] but is unrealistic for clinical decision making, because of the time required for an antigenic response.
The most frequently used method for Legionella detection is the urinary antigen assay, which was introduced shortly after the outbreak of legionnaires disease in Philadelphia [21]. This test revolutionized the laboratory detection of Legionella species, because it shows a high specificity, has rapid results, is readily available in simple-to-use commercial kits, has reasonably good sensitivity, and shows positive results even several days after initiation of therapy [15]. The test detects L. pneumophila serogroup 1, which accounts for 70%–80% of community-acquired cases of legionnaires disease [15, 16]. In a review of the results of a particular urinary antigen assay (Binax EIA) for 317 culture-proven cases, the sensitivity of the test was 94% for travel-associated legionnaires disease, 87% for CAP, and 44% for nosocomial legionnaires disease [16]. As with other tests for Legionella species, the yield of urinary antigen testing varies with the clinical severity of disease [22], and the yield is increased with concentrated urine samples [15].
C. pneumoniae. This is the atypical pathogen that has proved elusive in attempts to detect it. The most common method used in studies of CAP is serological testing, with a variety of techniques. Microimmunofluorescence serological analysis is the test that is considered standard, according to recommendations from the Centers for Disease Control and Prevention [23]. The criteria for diagnosis of acute infection is an IgM titer of 1:16 or greater or a 4-fold increase in IgG titer. The issues with the test are the liberties taken in compliance with these criteria and problems with the test itself. The concerns about microimmunofluorescence serological analysis are the false-positive and false-negative test results, attributed to delayed and unpredictable development of IgM and IgG, and a lack of standardized methods. One report indicated that microimmunofluorescence analysis showed serological evidence of acute infection in 19% of healthy adults [24]. With regard to false-negative results, one study reported a positive result for only 25% of children with culture-proven C. pneumoniae infections [25], and another reported uniformly negative results of microimmunofluorescence testing for 17 patients with CAP who had positive results of PCR assay [26]. There is also poor concordance among tests done in different laboratories [27].
There are no commercially available tests for detection of C. pneumoniae that are FDA cleared, but a number of laboratories use “in-house” PCR assays [28]. In 2001, the Centers for Disease Control and Prevention reviewed 18 PCR assays and considered 4 to be adequately validated [3]. A multicenter study of C. pneumoniae detection in atheroma samples that included spiked control samples showed substantial variation in results among 15 laboratories [29, 30].
The conclusions are that there is no commercially available test that is FDA cleared for detection of C. pneumoniae, that the most frequently used tests show substantial variation in technique, and that results often show poor concordance among laboratories or tests. Regarding diagnostic testing for therapeutic trials, it seems that retrospective analysis using standardized tests is easily done for Legionella species, is probably doable for M. pneumoniae, and is unrealistic for C. pneumoniae.
Prevalence of Atypical Pathogens in CAP
The frequency of these 3 pathogens among CAP cases varies among different studies on the basis of the methods used for detection, as well as the population studied. Nevertheless, most studies of CAP show that 20%–30% of cases will involve at least one of these pathogens [6, 31, 32]. There is substantial variation in the site of care; studies involving outpatients show high yields of M. pneumoniae and somewhat lower yields of C. pneumoniae, whereas studies involving patients with CAP who require admission to the intensive care unit show almost exclusively Legionella species. The University of Louisville Atypical Pathogen Reference Laboratory uses a variety of techniques for detection of atypical pathogens and has reported the results for specimens obtained in 1996–2004 from 4337 patients in 21 countries. The results showed an incidence of 21%–28% in adults with CAP in virtually all regions of the world (table 1) [33]. Other studies have sometimes shown similar yields, but generally there is great variation. Of possible importance is the observation that PCR to detect C. pneumoniae has shown very low yields, and it is not known whether this reflects inadequate sensitivity of the test, the population studied, the season of the study, or the possibility that this is not a common pathogen [32, 34, 35].
Table 1
Incidence of atypical pathogens in community-acquired pneumonia (CAP).
Is It Necessary to Treat for Atypical Pathogens in CAP?
On the basis of a literature review, there have been 3 meta-analyses performed to address this issue [36–38]. All 3 used the same study format with a review of comparative trials of agents showing activity against atypical pathogens (macrolides, fluoroquinolones, or tetracyclines) versus β-lactams. In 2005, the Cochrane Library Report [36] performed an analysis of 24 trials that included 5015 randomized patients. The results showed no significant difference between the group receiving an agent active against atypical pathogens and the group receiving β-lactams, in terms of survival or clinical efficacy. However, a subset analysis showed a statistically significant benefit of treatment for patients with infections caused by L. pneumophila. The review by Mills et al. [38] showed similar results—that is, no improvement in clinical outcome with the use of antibiotics active against atypical pathogens, except among the subset of patients infected with L. pneumophila (table 2). The most recent report [37] reviewed studies from 1966 through July 2007 involving outpatients with CAP and concluded that there was no advantage to using specific antibacterials for treatment of mild CAP in otherwise healthy outpatients. A potential concern about these reports is that they tended to analyze the same studies. Nevertheless, it would be fair to conclude that a substantial number of randomized trials have been performed to address the issue of treatment for atypical pathogens in CAP, and they do not provide compelling evidence of a need to treat atypical pathogens, with the exception of treatment for legionnaires disease.
Table 2
Meta-analysis comparing antibiotic coverage with no coverage of atypical pathogens in community-acquired pneumonia.
Studies addressing the individual pathogens are also of interest. Data for Legionella species seem uncontested, because the need to treat infection with these species was demonstrated very early, on the basis of anecdotal experience with antibiotics that showed a better survival rate among those patients who received agents that, in retrospect, were shown to be active against L. pneumophila [39]. Multiple reviews done 30–50 years ago seemed to show a clear response to antibiotics [40–44], including in a placebo-controlled trial [43] and a randomized trial that compared arithromycin, tetracycline, and penicillin [44]. Nevertheless, a recently reported randomized trial involving 84 children aged <5 years with established Mycoplasma CAP showed no difference in outcome for treatment with levofloxacin versus treatment with a β-lactam [45].
Support for atypical coverage is implied by the Medicare database, in which >13,000 hospitalized patients with CAP were analyzed for outcome on the basis of antibiotic selection and were stratified by severity of illness [46]. This database shows a statistically significant survival advantage with the use of a fluoroquinolone or a β-lactam plus a macrolide, compared with a β-lactam alone. This result supports the benefit of coverage of atypical pathogens, but the conclusion is limited on the basis of the uncertain value of these antibiotics, in terms of anti-inflammatory properties; the lack of any accompanying microbiological data; and the anecdotal nature of the data.
Conclusions
On the basis of controlled trials, there can be no consensus regarding the need to treat atypical pneumonia in all patients; however, controlled trials from 3–5 decades ago show that CAP caused by M. pneumoniae should be treated on the basis of clinical response, as demonstrated by several well-accepted clinical parameters, and that legionnaires disease must be treated, because antibiotics are clearly effective in vitro and in vivo and because the mortality rate with such treatment is generally reported to be 5%–10%, which is less than the mortality rate for no treatment. Conclusions about C. pneumoniae infections are limited by the problems associated with diagnosis. Randomized trials of agents that are active against atypical pathogens versus β-lactams would be difficult to conduct in countries that have guidelines that recommend routine coverage of atypical pathogens, which include the United States, Canada, the United Kingdom, Germany, and Japan. Countries in which atypical coverage is optional include France, Hong Kong, Saudi Arabia, and South Africa. Controlled trials in these regions seem feasible, because they would comply with standards of care.
Acknowledgments
Supplement sponsorship. This article was published as part of a supplement entitled “Workshop on Issues in the Design and Conduct of Clinical Trials of Antibacterial Drugs for the Treatment of Community-Acquired Pneumonia,” sponsored by the US Food and Drug Administration and the Infectious Diseases Society of America.
Potential conflicts of interest. J.G.B. is on the policy advisory board for Johnson and Johnson and is on the infectious disease advisory board for Pfizer.
- © 2008 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
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
