|Home | About | Journals | Submit | Contact Us | Français|
The date 2 November 2009 marked the first World Pneumonia Day, launched by a coalition of child health organizations to support global efforts to prioritize pneumonia treatment and prevention. Despite recent medical advances, pneumonia takes the lives of up to 2 million children under age 5 each year—more than AIDS, malaria, and measles combined (15). As with many other illnesses, the disease burden is greatest among the world's most vulnerable population: children living in developing nations. For every child who dies of pneumonia in a developed country, more than 2,000 die in developing countries (16).
With safe and effective vaccines against Haemophilus influenzae type b (Hib), Streptococcus pneumoniae, and measles virus and with improvements in environment and nutrition and in case management standards, we now have several strategies to both prevent and treat pneumonia. The challenge is to ensure universal access to these life-saving interventions.
Laboratory diagnostics and clinical microbiologists should play an important part in global efforts to prevent and treat pneumonia. Diagnostic testing has an essential role in ensuring the most appropriate and effective therapy for individual patients. It also plays a role in disease surveillance in defining the etiologic spectrum of pneumonia cases and deaths. This in turn forms the evidence base for strategic decisions by global decision makers (such as the World Health Organization, UNICEF, and the GAVI Alliance), vaccine manufacturers, and funders to develop and support treatment algorithms, vaccine products and programs, and other effective prevention strategies. World Pneumonia Day also reminds us that much work remains to be done in pneumonia diagnostics and that historically, even with the best of methods, we have been unable to define the microbial etiology of a significant proportion of pneumonia episodes, particularly in children. The conditions under which pneumonia mortality is greatest are also the conditions under which adequate diagnosis is least possible. Even in settings with access to state-of-the-art microbiological diagnosis, establishing the etiology of a pneumonia case is fundamentally vexed by the limited ability to obtain specimens from the site of infection without contamination by upper respiratory secretions. Lack of sensitive laboratory diagnostic tools has probably played a direct role in the delayed introduction of effective vaccines to prevent pneumonia. Vaccines against Hib and S. pneumoniae alone may prevent >50% of severe childhood pneumonia, but poor diagnostics contribute to substantial underdiagnosis, and thus, substantial efforts need to be put into place to build awareness about the potential impact of these interventions. In spite of these limitations, significant advances have been made in diagnostic technology. However, technology alone is not a panacea for pneumonia diagnostic needs: assessment of performance within clinical and epidemiologic contexts is essential.
The routine microbiological evaluation of patients with suspected pneumonia continues to rely on methods that have been around since Margaret Pittman's time: microscopy and culture of lower respiratory tract specimens, blood cultures, detection of antigens in urine, and serology. While there have been refinements in these traditional laboratory tools, they have led to only modest improvements in overall diagnostic capability. Indeed, some have argued that we are doing less well now in diagnosing the cause of pneumonia than we were doing 70 years ago (2).
Recent advances in pneumonia diagnostics have most notably occurred in the areas of antigen and nucleic acid detection. Commercial antigen detection assays are now widely available for several pneumonia pathogens, particularly S. pneumoniae, Legionella pneumophila, and some respiratory viruses. A new generation of immunochromatographic pneumococcal urinary antigen tests that detect the C-polysaccharide cell wall antigen have proven useful for diagnosing pneumococcal pneumonia in adults (19). Detection of soluble Legionella antigen in urine is now well established as a diagnostic tool for Legionnaires' disease (6). Several commercial rapid tests, using immunochromatography, enzyme-linked immunosorbent assay, or other formats, are now available for respiratory viruses, including influenza and respiratory syncytial viruses (11, 12). The ease of performance of these assays has been responsible for their widespread use, including their use as near-patient tests.
Nucleic acid detection tests (NATs), such as PCR, have been developed for all major pneumonia pathogens, although the development of commercial assays and their introduction into routine use by diagnostic laboratories have been relatively slow (7). NATs have many features that make them attractive tools for diagnosing the etiology of respiratory tract infections. They detect very low levels of nucleic acid from potentially all respiratory pathogens, do not depend on the viability of the target microbe, can provide results within a clinically relevant time frame, are probably less affected by prior antibiotic administration than culture-based methods, and have the potential to provide supplementary information, such as the presence of antibiotic resistance genes. NATs have been particularly useful for diagnosing infections that are difficult to rapidly diagnose by other methods (e.g., rhinovirus infection) and are now regarded as the preferred methods for detecting some respiratory pathogens, such as the respiratory viruses and Chlamydophila pneumoniae.
Despite these technological advances, there are several major challenges that hinder the search for and understanding of the etiology of pneumonia.
The detection of known pathogens in good-quality specimens collected directly from the lower respiratory tract provides good evidence of the likely microbial etiology of pneumonia, especially for microorganisms that do not normally colonize the upper respiratory tract. However, obtaining these specimens can be difficult, creating a fundamental problem with pneumonia diagnostics. While efforts are directed toward the development of assays with ultrahigh analytical sensitivity, the fact that an appropriate clinical specimen may not even be obtained from a patient in the first place is often overlooked in research activities. Even in adults with pneumonia, a large proportion do not produce sputum, including less than half of those with Legionnaires' disease (6). Obtaining sputum from children is much more difficult and is usually possible only through induction procedures, which may in themselves cause concern about safety in children with respiratory distress. Difficulties obtaining sputum may prompt the use of more-invasive techniques, such as bronchoscopy and transthoracic needle aspiration (18). The latter is a relatively safe procedure if it is performed by trained staff on appropriately selected patients and has the potential to define pneumonia etiology with high sensitivity and specificity. Further efforts to enhance the collection of specimens from the lower respiratory tract could make a major contribution to improving the identification of etiologic agents from pneumonia episodes.
There are two particular challenges when testing respiratory specimens in the context of pneumonia. First, for a pneumonia pathogen that can colonize the upper airways of healthy individuals (e.g., S. pneumoniae), does detection in a lower respiratory specimen represent infection or contamination with a colonizing organism? Second, for a pathogen that replicates in (but does not colonize) the upper airways and causes a spectrum of diseases (e.g., rhinovirus), does detection in a respiratory specimen indicate that this organism is the cause of pneumonia or could this be a coincidental finding following a recent unrelated upper respiratory tract infection (URTI)?
Distinguishing colonization from infection is a major challenge when processing any sputum specimen. Quality checks based on the relative numbers of squamous epithelial cells and leukocytes seen in a sputum Gram stain smear may help detect upper airway contamination (8), especially if correlated with culture results. Despite the ongoing debate about the utility of sputum culture, there is at least some evidence that prompt examination of a high-quality sputum specimen collected before antibiotic treatment has reasonably high sensitivity and specificity for pneumococcal pneumonia in adults (9). Lower respiratory tract specimens obtained by bronchoscopy can also be contaminated with upper airway flora, but the likelihood of contamination is generally regarded to be less.
All the major viruses that cause pneumonia are more usually associated with nonpneumonic illnesses, particularly URTIs. Viral URTIs are common, especially in young children who typically experience >4 URTIs annually. With each episode, virus can typically be detected for at least 7 days, which means that the average child may test positive for a respiratory virus for at least 1 month of every year. Some viral URTIs occur at the same time as pneumonia, and these may be causal (either directly or by leading to secondary bacterial pneumonia) or may be incidental. In a study setting, these can be distinguished by determining the background prevalence of nasopharyngeal viral infection in a control group. Unfortunately, most studies of viral pneumonia have not used control groups and have therefore not questioned the implication that a positive test from a nasopharyngeal specimen indicates causation (5, 14). To further complicate matters, there is the problem of how to ascribe causality when more than one pathogen is detected within an individual; this is more likely to occur with use of multiple sensitive testing methods.
Distinguishing colonization from infection extends to the testing of nonrespiratory specimens as well. Although newer-generation pneumococcal urinary antigen tests have been an important advance for adults, they cannot be used reliably with children since they detect pneumococcal carriage in this age group (4).
Most diagnostic tests for pneumonia pathogens have suboptimal diagnostic sensitivity. Blood cultures are frequently performed for hospitalized pneumonia patients but are positive in <10% of cases. Blood cultures are also the main diagnostic tool used in surveillance programs for invasive pneumococcal disease and are clearly able to identify only a minority of patients with pneumococcal pneumonia. Some fastidious pneumonia pathogens, such as Legionella spp. and C. pneumoniae, can be difficult to isolate even with the use of special culture methods. Current commercial urinary antigen detection assays can reliably detect only infection caused by L. pneumophila serogroup 1 and not other species and serogroups. Serological testing may return falsely negative results if the convalescent-phase serum is collected too early, before a detectable increase in antibodies, or not collected at all or if acute-phase serum is collected too late. While having the advantage of producing quick results, rapid antigen tests for influenza virus and respiratory syncytial virus are generally less sensitive than culture-based methods.
Unfortunately, many new tests (particularly NATs) are introduced into diagnostic laboratories before adequate sensitivity and specificity data are generated. Before a test is introduced into a clinical diagnostic laboratory, it should first be thoroughly evaluated to determine analytical sensitivity and specificity. It should then be applied to different populations to determine clinical sensitivity and specificity, which are distinct from analytic sensitivity and specificity measures (10). If the test is to be applied to multiple specimen types, clinical sensitivity and specificity need to be determined for each specimen type.
All too often, analytical sensitivity is confused with clinical (diagnostic) sensitivity (10). Analytical sensitivity (lower limit of detection) refers to the smallest amount of target that can be accurately detected by the assay, is relatively straightforward to calculate in laboratory-based studies, and is expressed as a concentration. Clinical sensitivity is the proportion of true-positive patient samples that are correctly identified by the assay and is a measure of test validity in a population. Its measurement involves testing clinical specimens and comparing the new test with a “gold standard” test; it is expressed as a proportion or percentage. In spite of these important differences, the two terms are frequently used interchangeably in clinical settings. Very high analytical sensitivity is often used to promote NATs, despite the fact that high analytical sensitivity does not guarantee high clinical sensitivity (e.g., as occurs when blood-based inhibitors of Taq polymerase are not removed prior to annealing). Although it is desirable for an assay to have a high analytical sensitivity, the biological relevance beyond a certain point may be minimal. An assay that is so sensitive that it detects harmless asymptomatic infection may lead to difficulties in attributing etiology to a pathogen because of the frequency with which it is found among control subjects. In these circumstances, a quantitative approach may be preferable.
Another difficulty is the accurate calculation of clinical specificity due to the lack of suitable “gold” standards. This is an inherent problem with all NATs, both within and outside the context of pneumonia. NATs are likely to have higher clinical sensitivity than most traditional diagnostic tools. Consequently, a positive NAT result cannot necessarily be dismissed as being falsely positive simply because a less-sensitive comparison test is negative.
Tests with high clinical sensitivity and specificity do not entirely address the issue of causality. A pathogen may truly be present in a patient as evidenced by a reliable assay, but that does not, for every pathogen, allow us to ascribe a given episode of disease to that pathogen. Epidemiologic sensitivity and specificity, which can be established only from population-based studies which include control subjects and employ a panel of highly sensitive and specific assays, inherently include the analytic, clinical, and causal components of pathogen detection. Vaccine probe studies, which uncover the attributable burden of a given pathogen, are another approach to firmly ascribing the proportion of pneumonia to one or more pathogens.
To advance pneumonia diagnostics, we need to improve current technology and simultaneously explore innovative approaches. We need to give attention to the specimen type and how we collect it equal to that given to laboratory assay development. New diagnostics must be systematically and rigorously evaluated for analytic, clinical, and eventually epidemiologic sensitivity and specificity. Strategies to determine the etiology of pneumonia in children are a special priority.
NATs have now been developed to a stage whereby multiplex assays that detect all common respiratory pathogens are commercially available in user-friendly platforms, and further improvements in design and performance are expected. The emphasis should now be placed on clarifying the clinical usefulness of these assays, developing standardized methods, producing even more user friendly platforms, and exploring the role of quantitative assays. Antigen detection assays in immunochromatographic or similar formats are fast and simple to perform. In many ways these methods are among the most attractive diagnostic tools, as illustrated by the success of rapid diagnostic tests for malaria (17), but further development is dependent on the discovery of suitable antigens that can be reliably detected in readily obtained samples.
Breath analysis is an example of an innovative area with enormous potential for pneumonia diagnostics. Bacteria and fungi produce volatile metabolites that may be used as biomarkers (1). Detection of these biomarkers in breath samples by gas chromatography/mass spectroscopy or similar methods may provide an etiological diagnosis of pneumonia. Potential biomarkers have been reported for some respiratory pathogens (3, 13), but it is still uncertain whether they will prove to be useful as clinical diagnostic tools. Regardless, this sort of innovation is welcome in the world of pneumonia diagnostics.
Last, it is important to remember that the greatest burden of pneumonia is in developing countries. To enable deployment in underresourced settings, laboratory diagnostic tests for pneumonia need to be rapid, inexpensive, and easy to use and should be useful for surveillance purposes. However, there is also a role for more-expensive high-tech tests that may not require wide availability. There are good reasons to determine the causes of pneumonia which may be best satisfied by deployment of sophisticated tests in sentinel sites, viz., (i) development of empirical treatment guidelines in the world following the introduction of vaccines against Hib and S. pneumoniae and (ii) evaluation of the impact of vaccines (and other preventive interventions) on the incidence of pneumonia. As the spectrum of pneumonia-causing pathogens changes because of vaccine introduction, antimicrobial resistance, HIV/AIDS, urbanization, and socioeconomic changes, our ability to prevent and treat this life-threatening syndrome is anchored in our ability to monitor the etiologic spectrum of disease-causing pathogens. Significant efforts will be required in these settings to build and sustain the capacity of local laboratories and laboratory workers and to link laboratory findings with clinical management of individual patients. World Pneumonia Day provides a useful reminder that much more can and must be done to improve the diagnosis, treatment, and prevention of pneumonia worldwide and that clinical microbiologists play an important role in the process.
This work was performed under the Pneumonia Etiology Research for Child Health (PERCH) program at the Johns Hopkins Bloomberg School of Public Health and was funded in full by a grant from the Bill & Melinda Gates Foundation. N.B. is supported by grant 1KL2RR025006-01 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research.
The contents are solely our responsibility and do not necessarily represent the official view of NCRR or NIH. Information on NCRR is available at http://www.ncrr.nih.gov/. Information on Re-engineering the Clinical Research Enterprise can be obtained from http://nihroadmap.nih.gov/clinicalresearch/overview-translational.asp.
The views expressed in this Commentary do not necessarily reflect the views of the journal or of ASM.
Published ahead of print on 9 September 2009.