The global challenge of bacterial pneumonia

Antibiotic-resistant bacteria is shaping up to be the clinical challenge in the 21st century.¹ It was only 10 years ago that the Infectious Disease Society of America (IDSA) published an article titled “Bad Bugs, No Drugs,” warning of the rise of threatening microorganisms coupled with stagnation of new antibiotic drug discovery.² Unfortunately, the superbug and drug resistance situation has grown considerably worse over the last decade. Apart from some new therapeutic options for MRSA, there is a paucity of novel antibiotics for treating resistant pathogens like Klebsiella pneumonia, Acinetobacter baumannii, and Pseudomonas aeruginosa. These bacteria play a key role in many severe infections, particular in pneumonia—a life-threatening problem in acutely ill patients. Moreover, there is an increasing incidence of mixed infections and a growing percentage of multidrug-resistant bacteria, and both of these factors are directly associated with the risk of empiric treatment failure. Given the lack of new antibiotics, this development is threatening the ability to effectively treat serious infections. Several studies have clearly demonstrated that if inadequate antibiotic therapy is administered to patients, the mortality rate and mean duration of hospital stay will significantly increase.^3,4 Microbiology culture technology is currently the standard of care in diagnosing pneumonia. However, culturing is a slow process and can delay treatment of patients with the appropriate targeted antibiotic therapy. New molecular methods such as multiplexed PCR assays may overcome current challenges in handling severe pneumonia cases.

The scope of the problem

The respiratory tract is the most common source of infection caused by viruses, bacteria, or fungi. Recent data shows that bacterial co-infection complicated one in four patients admitted to ICU with severe influenza.⁵ Viral upper respiratory infections can become life-threatening when associated with a bacterial infection turning into an infection of the lung (pneumonia). 

In general, the frequency of pneumonia ranged from 64% to 68% in the EPIC II⁶ and SOAP⁷ studies, respectively. Pneumonia is categorized according to the location of onset: community-acquired pneumonia (CAP) and hospital-acquired pneumonia (HAP) are the major, well-defined categories. HAP can be further divided into ventilator-acquired (VAP) and non-ventilator acquired pneumonia. Five to 10 cases of HAP are diagnosed in every one thousand patient admissions into the hospital. They are associated with an additional 12 days of ventilation, doubled mortality, and prolonged intensive care unit (ICU) length of stay by six to 19 days per patient. Staphylococci and Enterobacteriaceae are more prevalent in HAP and VAP, with P. aeruginosa and A. baumannii becoming more common where patients have been hospitalized or ventilated for a longer period of time (late-onset pneumonia). The total average hospital cost for a HAP patient is approximately $21,000. These costs rise to more than $67,000 in patients with VAP, due to a greatly increased mean length of stay in the ICU. 

This situation has become even more complicated as Gram-negative infections are increasing worldwide, accounting for the majority of infections in ICUs (63% in 2007 versus 39% in 1995). In the case of Gram-negative infections, P. aeruginosa was the most frequent organism found in infected ICU patients (20%), followed by K. pneumonia (13%) and A. baumanii (9%). The high values for P. aeruginosa infections are troubling because of its difficult treatment and association with increased mortality in ICU patients. In addition, the trend of growing antibiotic resistance is continuing to worsen with the rise of globally successful resistant lineages such as the E. coli ST131 and others. The organisms of greatest concern include ESBL-positive Enterobacteriaceae, carbapenem-resistant Enterobacteriaceae, carbapenem-resistant A. baumanii, and MBL-positive P. aeruginosa. Two cases demonstrate this dangerous evolution:

 From 2005 to 2007, Greece dealt with an epidemic of carbapenem-resistant VIM-producing K. pneumoniae. This was immediately followed by a new epidemic of K. pneumoniae, expressing KPC type beta-lactamases.⁸

 Until 2000, most extended spectrum beta-lactamines (ESBLs) were mutants of classical TEM and SHV plasmid-mediated penicillinases.These were found mostly in Klebsiella and rarely in E. coli. Since 2000, the greater problem has become the proliferation of cefotaximase-Munich (CTX-M)-type ESBLs also increasing in the US.⁸

In addition, mixed infections are becoming a big concern, as the SOAP trial demonstrated the presence of a polymicrobial infection in 23% of cases observed. Another study showed that in patients infected with a single pathogen, 90% received appropriate empiric antibiotic therapy. By contrast, in patients diagnosed with more than one pathogen, nearly 40% were inappropriately treated.⁹

Empiric antibiotic therapy with wide coverage is typically given in generous doses. However, with the growing incidence of Gram-negative pathogens, drug resistance, and mixed infections, this current standard of care might fail, and even support the development of new resistances. Once the bacteriological data are available, the antibiotic treatment spectrum should be narrowed and adjusted based on the findings. Since microbiology culture techniques are very slow, in most cases these data are provided too late in the decision process. Because of the limitations of culture, published treatment guidelines might be a useful tool in selecting empirical antibiotic therapy. However, since they are based on scientific evidence and expert recommendations, it is important to validate these recommendations in the clinical setting. A study of HAP patients treated in the ICU was conducted to make this determination. Two patient groups were selected according to the guidelines; group one consisted of patients without an antibiotic resistance risk, while group two enrolled patients with risk factors for resistant microorganisms. Microbial prediction was poorer in group one (55% correct) than in group two (82% correct). The major cause for this discrepancy was the presence of microbiologically confirmed drug-resistant microorganisms in patients in group one, which were wrongly assessed by using the guideline-based classification—a result clearly demonstrating the limitations of relying solely on treatment guidelines.¹⁰

Diagnostic options

For diagnosing influenza, rapid molecular testing, even multiplexed testing, has been commercially available and in routine use for a couple of years, e.g., xTAG RVPv1, ResPlex II, MultiCode-PLx. However, the identification of non-viral respiratory pathogens in a patient with pneumonia is currently a slow process. The standard of care generally requires 18 to 24 hours to grow the pathogen(s) from a sample and an additional six to 48 hours to identify the pathogen and to conduct antibiotic sensitivity testing. Moreover, classical culture methods are not always reliable since some pathogens, such as Mycoplasma pneumoniae, are hard to culture, and some resistance pathways, like ß-lactamases, offer only limited detection. Classic microbiology methods include serology, in which seroconversion or immunoglobulin titers are measured to confirm the presence of organisms such as L. pneumophilia with sensitivities only in the range of 70% or lower. 

In the hope of speeding the process of determining resistance from culture, technologies have been developed that incorporate both biochemical pathogen identification and susceptibility testing in an automated system. In addition, IDV companies have focused on automated plate reading, which improves the efficiency of the laboratorian. Some have been successful in accelerating the time to pathogen identification with mass spectrometry to within 15 minutes. However, most of today’s approaches do not circumvent the need to culture the microorganism and thus do not significantly impact the timeline to deliver results to the clinician.

Moreover, to improve turnaround time for pathogen identification, new methods should address the challenge that clinical samples are heterogeneous. For instance, respiratory tract specimens come from a broad range of fluids such as sputum, tracheobronchial aspirate, and bronchoalveolar lavage as well as pleural puncture. 

Methods of pathogenic testing should also recognize prevalent drug resistances, which is already in routine use for some infections. In the case of MRSA, methicillin resistance is predominately determined by the acquired mecA gene, which can easily be detected by molecular methods. A similar example is carbapenem resistance in K. pneumoniae, which involves the genes encoding KPC, IMP, VIM, NDM, and OXA 48 carbapenemases, and can be assessed by PCR. This is also true for quinolone resistances in Enterobacteriaceae, as they are largely mutational. However, resistance mechanisms are becoming increasingly complex and dynamic as gene acquisitions, mutations, and mosaic genes are all important factors driving antibiotic resistance.

The general pattern is that resistance mostly arises from gene acquisition in Enterobacteriaceae, Acinetobacter, and S. aureus. Mosaic genes are important in transformation-competent species such as pneumococci, whereas gene mutation is the predominant mechanism in P. aeruginosa. Although any phenotype depends on the genotype, the correlation may be variable due to the level of expression of the resistance gene product, the presence of cofactors expressed along with the resistance gene, and the possibility that different resistance genotypes result in common resistance phenotypes. For instance, E. coli isolates producing TEM-1 are resistant to ampicillin and amoxicillin, but are normally susceptible to amoxicillin plus clavulanic acid. Resistance to amoxicillin-clavulanic acid can be due to overproduction of TEM-1, providing an example of how the same resistance genotype may be associated with different phenotypes.¹¹ 

These types of findings underscore the potential limitations of genotypic analysis. Nevertheless genotyping can be more reproducible and robust than phenotyping in certain instances. One study that compared the results of susceptibility testing of VIM-1-producing K. pneumoniae isolates observed that results of carbapenem susceptibility exhibited remarkable discrepancies when using different assays.¹² Another example is the genotyping of carbapenemase genes, which can be more reliable for detection of the presence of resistance determinants and assessment of the risk of a carbapenem-resistant phenotype. 

Molecular diagnostic approaches

Today’s physicians have access to a number of new tests and technologies recently introduced into the market that have the potential to bring automated diagnostic approaches for pneumonia into clinical routine. Although the problem of rapid pathogen identification has been mostly solved, the need for initial culture to perform antibiotic susceptibility testing is still the rate-limiting and often costly step. With the rising problem of drug resistance, clinical microbiologists are often faced with the need to carry out supplemental susceptibility testing, including the D-test, the ESBL-confirmatory test, the modified Hodge test, and the metallo ß-lactamase test. These tests can add another 18 to 24 hours to the overall time to initial diagnosis. Therefore, to identify respiratory pathogens in conjunction with antibiotic resistance markers rapidly within hours, novel approaches are still desperately needed.

A few solutions, such as the GeneXpert, FilmArray, Verigene, or Unyvero*, are providing molecular diagnostic approaches. By incorporating a multiplexed PCR-based analysis, those technologies can cover a broad range of analytes including Gram-negative bacteria, while also combining pathogen identification with the detection of resistance markers in a single automated procedure within a few hours. Some of these technologies have improved sample preparation, allowing the use of native specimens of any origin that do not require culture. European study data shows that turnaround time using solutions such as Unyvero is approximately five hours, while conventional culture-based microbiology results averaged 49.7 hours for pathogen identification.¹³ In addition, the combination of conventional PCR and microarray technologies in a cartridge-based sample-to-answer system allows the detection of complex antibiotic resistance mechanisms, as up to 100 analytes can be detected simultaneously.  

The manufacturers of the aforementioned new systems stress their ease of use and that they can be performed closer to the patient, with the aim of speeding current microbiology procedures typically carried out in large, centralized clinical laboratories. They assert that automation and simplification of molecular diagnostic platforms will also facilitate adoption of molecular testing platforms in contexts in which specially trained microbiology staff are in short supply. They say that the successful implementation of rapid molecular diagnostic testing can improve turnaround time from days to hours and have a significant positive impact on patient management and outcome.

To sum up, bacterial pneumonia is associated with significant mortality and high economic costs. The ultimate goal of a pneumonia therapy is to apply the appropriate, most effective antibiotic regimen as early as possible. However, antibiotic testing based on current microbiology detection methods is often inadequate. The growing incidence of Gram-negatives, increasing presence of mixed infections, and the higher probability of multidrug-resistant pathogens all further complicate effective, timely pathogen detection and successful treatment. The availability of rapid molecular diagnostic offerings could result in more effective antibiotic therapy by increasing accurate diagnoses and therefore the percentage of pathogen-directed targeted treatments. Some emerging molecular diagnostic technologies not only hold the potential to improve patient care but could also help decrease antibiotic resistance in the general population, paving the way to responsible antibiotic stewardship by prolonging the effectiveness of existing drugs. However, the introduction of new diagnostic tests in patient care is a remarkable challenge in times of cost-cutting in healthcare budgets. Therefore, more studies and randomized trials to evaluate the medical and economic benefits of these rapid molecular tests are needed. As more clinical data becomes available, it will also be possible to accurately evaluate the clinical impact of genotyped antibiotic-resistance testing on patient outcomes.