New Microbiological Techniques in the Diagnosis of Bloodstream Infections

Abstract

Background

When a bloodstream infection is suspected, the preliminary and definitive results of culture-based microbiological testing arrive too late to have any influence on the initial choice of empirical antibiotic treatment.

Methods

This review is based on pertinent publications retrieved by a selective search of the literature and on the authors’ clinical and scientific experience.

Results

A number of technical advances now enable more rapid microbiological diagnosis of bloodstream infections. DNA-based techniques for the direct detection of pathogenic organisms in whole blood have not yet become established in routine use because of various limitations. On the other hand, matrix-assisted laser desorption/ionization—time of flight (MALDI-TOF) mass spectrometry (MS) has become available for routine use in clinical laboratories and has markedly shortened the time to diagnosis after blood samples that have been cultured in automated blood-culture systems turn positive. Further developments of this technique now enable it to be used directly for blood cultures that have been flagged positive, as well as for subcultures that have been incubated for only a short time on a solid nutrient medium. The microbial biomass of the subculture can also be used in parallel for more rapid susceptibility testing with conventional methods, or, in future, with MALDI-TOF MS.

Conclusion

The potential of all of these new techniques will only be realizable in practice if they are optimally embedded in the diagnostic process and if sufficient attention is paid to pre-analytical issues, particularly storage and transport times.

From the very beginning, the task of diagnostic microbiology was rapid, confident identification and characterization of pathogens. This is particularly true for bloodstream infections, which—especially in the case of sepsis—acutely endanger the patient (e1).

Rapid detection is especially important in light of the virulence of many pathogens and the existence of resistent phenotypes. Moreover, we are confronted by a population increasingly vulnerable to infection, with ever more elderly, multimorbid, and immunocompromised individuals together with growing numbers of patients with implants and other foreign bodies (e2).

Pathogen diagnosis.

Rapid diagnosis is important due to pathogen virulence and the existence of resistant pathogen phenotypes, and also because the population structure is increasingly vulnerable to infections, with ever more elderly, multimorbid, and immunocompromised patients and/or those with implants and other foreign bodies.

Polymerase chain reaction (PCR).

With the development of the PCR technique, pathogen detection was accelerated and higher sensitivity and specificity were achieved for pathogens that could be cultured only with difficulty or not at all.

The rapidity of pathogen detection in culture is limited by the rate of division of the microorganisms concerned. Traditionally, there were “overnight” cultures for each culture-based step of diagnosis (primary culture, isolation of mixed cultures, biochemical identification, and susceptibility testing), often resulting in diagnosis times of 2 to 3 days or longer. The advent of nucleic acid amplification tests (NAT) in the form of the polymerase chain reaction (PCR) therefore amounted to a revolution: pathogen detection was accelerated, and higher sensitivity and specificity were achieved for the detection of fastidious or non-cultivable pathogens (1, e3, e4). More recently, mass spectrometry (MS) techniques have again revolutionized pathogen identification and in many institutions already form part of the routine diagnostic work-up. Further far-reaching changes can be expected from whole genome/microbiome sequencing and from automation of culture-based procedures.

Learning goals

After reading this article, the reader should be:

• Familiar with the essential aspects of recent and coming developments in pathogen detection

• Able to assess both the advantages and the disadvantages of these procedures

• Acquainted with the use of recently developed techniques as they apply to the diagnosis of bloodstream infections

Methods

This review article is based on a selective literature survey and on our own diagnostic and research experience in the use of NAT and matrix-assisted laser desorption/ionization—time of flight (MALDI-TOF) MS for microbiological diagnosis of bloodstream infections.

Principles and overview

PCR and culture techniques.

Depending on the indication, PCR and culture can contribute synergistically to valid identification of the pathogen, determination of its susceptibility, and definition of further characteristics.

Nowadays PCR, other NAT, and sequencing techniques belong to the standard repertoire of microbiological diagnosis. These procedures and developments thereof are widely used for direct detection of pathogens or for molecular identification of cultured microorganisms, for detection of resistance genes or virulence factors, and for genotyping (e5– e8). On the other hand, it is being increasingly appreciated that culture procedures are in themselves highly sensitive (one microbial cell capable of multiplying in culture suffices), and only a culture isolate permits determination of the phenotype, which is more than the sum of all resistance-coding genes or mutations. Innovations in culture-based diagnosis also serve to detect pathogens with greater speed, sensitivity, and specificity.

Molecular genetic tests and phenotypic characterization based on the cultured isolate both have method-inherent limitations, but both also possess particular diagnostic advantages. Depending on the indication, these two diagnostic strategies can contribute synergistically to valid pathogen identification and determination of susceptibility, as well as closer characterization of the pathogen.

With regard to its influence on the quality and speed of pathogen identification, the introduction of MS techniques to the routine microbiological diagnostic work-up is comparable with the “PCR revolution”. In only a short time, MALDI-TOF MS systems have almost completely replaced conventional biochemical identification of pathogens in many laboratories and are well on the way to being employed for antibiotic susceptibility testing (e9– e12).

Pathogen detection by means of nucleic acid diagnostic tests

MALDI-TOF mass spectrometry systems.

In only a short time, MALDI-TOF MS systems have almost completely replaced conventional biochemical identification of pathogens in many laboratories and are well on the way to being employed for antibiotic susceptibility testing.

After the introduction of PCR and other NAT (figure 2) in 1985, their lack of dependence on culture and their high sensitivity and specificity swiftly led to acceptance of nucleic acid-based diagnostic techniques (e3). Typically, these procedures achieve analytic sensitivity of 1 to 100 genome equivalents of bacterial or fungal DNA. Specimens that are complex or exhibit inhomogeneous distribution of pathogen (e.g., tissue, blood, and stool) may lead to drastic reductions in sensitivity and specificity. The specificity of NAT depends heavily on the choice of target structures, the composition of primers and probes, and the PCR/microarray reaction conditions. The obvious advantages of NAT often lead to thoughtless and insufficiently validated employment of these techniques (e13). Among other errors, the influence of pre-analytic factors, the importance of optimal nucleic acid extraction, and the risk of contamination are frequently underestimated (e14– e16). To avoid similar “traps” when using recently developed innovative techniques, it is essential to consider the findings in conjunction with those of other methods and view them critically (box 1).

Figure 2.

Specific and nonspecific pathogen detection by means of DNA-based methods

*¹ Framing taxon-specific sequence segments; *² or by alternative DNA-based methods; *³ currently in development

BOX 1. Questions on assessment of NAT findings.

• Can a positive NAT finding in itself explain the etiology of a suspected disease?

• Does a positive NAT result really reflect the pathogen of the suspected infectious disease, or may it represent contamination of the specimen, e.g., by microorganisms that colonize the skin?

• May the NAT findings contradict the results of culture-based or other investigations?

• Can inappropriate pre-analytic handling (sampling and transport) be a cause of false-positive or false-negative results?

• Is detection of resistance genes by NAT of patient samples associated with the suspected infectious pathogen, or do the genes come from other microorganisms (e.g., detection of the mecA gene from coagulase-negative staphylococci also present in the sample)?

Specificity of nucleic acid amplification testing (NAT).

The specificity of NAT depends heavily on the choice of target structures, the composition of primers and probes, and the PCR/microarray reaction conditions.

If a given patient’s infection cannot be attributed to the presence of defined pathogens by means of conventional diagnostic tests and species-specific NAT, broad-range nucleic acid-based methods can be used (Figure 2, eSupplement 1) (2, 3, e16– e19).

Whole-genome sequencing (WGS) and whole-metagenome sequencing (WMS) procedures and so-called next-generation sequencing (NGS) techniques are close to being ready for routine application (eSupplement 1) (4– 6, e20– e22).

In view of the differences among various NAT systems with regard to detection range, potential limitations of applicability for different specimens, and quality criteria (sensitivity, specificity, predictive value), close cooperation is advisable between the clinicians and the microbiology lab in order to select the techniques best suited to meet the clinical requirements. This is particularly true when NAT procedures are used in addition to conventional diagnostic techniques.

Evaluation of results.

To avoid the danger of misinterpretation when using recently developed innovative techniques, it is essential to consider the findings in conjunction with those of other methods and view them critically.

MALDI-TOF MS

The use of MS for pathogen identification was first proposed in the 1970s, but not until later did the German biophysicists Franz Hillenkamp and Michael Karas develop a technique, MALDI-TOF MS, suitable for routine application (for explanation of the principles of this method, please see eSupplement 1) (e23, e24). MALDI-TOF MS breaks down microbial, mainly ribosomal, proteins, using the data thus acquired to identify bacterial and fungal microorganisms. The main advantages of the technique are its extreme rapidity (a matter of minutes) and the cost efficiency of individual analyses with high specificity. Although MALDI-TOF MS is a phenotypic procedure, its high specificity, almost equal to that of NAT, is due to the high proportion of ribosomal proteins in the spectrum generated, practically reflecting the ribosomal nucleic acid sequences in entirety.

Owing to improvements in analytical methodology, analysis software, and the convenience of the devices used, MALDI-TOF MS has gained swift acceptance in the field of clinical microbiological diagnosis, to the point where it is now the standard for the identification of bacteria and yeasts and has led to considerable changes in diagnostic routine (7, e25, e26). The shortening of pathogen identification times achieved by MALDI-TOF MS may, for example, lead to improvement in clinical outcomes (e.g., survival rates) (8– 10).

Premises of microbiological diagnosis of bloodstream infections

Bloodstream infections, including sepsis and particularly septic shock, represent one of the most difficult challenges to diagnosis and treatment. The advances made in the past 25 years have, however, transformed our understanding of the nature of sepsis and led to altered definitions (most recently the Third International Consensus Definitions for Sepsis and Septic Shock—Sepsis-3) (e1, e27). Microbiological methods for diagnosis are still not sensitive enough to confidently detect the pathogen in all cases of bloodstream infection, and in particular they remain too slow to be of use in deciding on the initial empirical antibiotic treatment. They are indispensable, however, for decisions on specific antibiotic treatment after pathogen identification and susceptibility testing, as well as indirectly in compiling a database for use in future empirical treatment. A modern microbiological diagnostic work-up must be judged by how far it increases the sensitivity for detection of bloodstream infections and by how close it can come to the so-called “golden hour” of initial antimicrobial therapy.

MALDI-TOF MS.

The main advantages of MALDI-TOF MS are its extreme rapidity (a matter of minutes) and the cost efficiency of individual analyses with high specificity.

Sepsis-related mortality can be greatly reduced by early initiation of adequate antimicrobial treatment. The first hour after onset of the symptoms is crucial for the prognosis (11). Although there are no randomized clinical trials with significant conclusions regarding mortality and length of hospital stay, more rapid microbiological diagnosis could be of considerable benefit for the treatment of bloodstream infections (12, e28– e30). Early studies have confirmed a significant influence of rapid pathogen identification and susceptibility testing on mortality, length of stay in critical care, and costs (e30).

It must be emphasized that both classic pathogen culture and molecular procedures, in themselves, do no more than demonstrate the presence of a microorganism (bacteremia or fungemia) or its DNA in a sample. The confident classification of a microbiologically detected organism as clinical pathogen of a bloodstream infection requires knowledge of the clinical context and acquaintance with all infection-relevant parameters.

Sepsis-related mortality.

Sepsis-related mortality can be greatly reduced by early initiation of adequate antimicrobial treatment. The first hour after onset of the symptoms is crucial for the prognosis.

Despite partial successes in reducing methicillin-resistant Staphylococcus aureus (MRSA) in particular countries or regions, the rates of multidrug-resistant pathogens are rising worldwide (e31– e35). For this reason, and in light of new commercially available antibiotics, (e36, e37), proper selection of antimicrobial treatment is decisive not only for the individual patient, but also for lowering selective pressure for the development of resistance.

The goal must be a swift transition to targeted treatment, based on the results of pathogen identification and susceptibility testing, in the interests of diagnostic and antibiotic stewardship.

Advances in the diagnosis of bloodstream infections

The conventional diagnosis of bloodstream infections consists of three steps:

• Inoculation of the patient’s blood into a liquid medium (blood culture bottles) for primary pathogen culture

• Subculture on solid growth media

• Identification and antimicrobial susceptibility testing (figure 3)

Figure 3.

Processes and times for classic pathogen diagnosis of bloodstream infections (gray) and modern means of accelerating diagnosis (white)

This procedure generally takes 3 to 4 days, perhaps longer in the case of compromised (previous antibiotic therapy) or slowly growing pathogens, and is therefore too slow to serve as a basis for early decisions on treatment (11).

For some decades blood culture bottles have been incubated in devices that automatically monitor microbial growth and emit a signal when a threshold value is attained (e38). For species determination and susceptibility testing, the sample is usually transferred from liquid growth medium to solid growth media for overnight incubation to produce individual colonies. In parallel to culture on solid growth media, a Gram stain is prepared and examined under the microscope (figure 3). The microscopy findings will ideally be delivered to the clinician 20 to 30 min after positive flagging of the blood culture and used to adjust the antibiotic regimen.

Up to a short time ago, the subsequent identification of individual colonies of microorganisms was based almost exclusively on biochemical methods, which required renewed microbial growth over a period of several hours and often could not be evaluated until after a further overnight incubation. Potential ways of accelerating the diagnosis of pathogens responsible for bloodstream infections rest either on more rapid pathogen identification from blood culture bottles flagged as positive or on NAT-based direct detection of the pathogen from blood samples as a complement to culture (figure 3).

Classification of findings.

The confident classification of a microbiologically detected organism as clinical pathogen of a bloodstream infection requires knowledge of the clinical context and acquaintance with all infection-relevant parameters.

Faster pathogen identification from positive blood culture bottles

DNA-based procedures

Recent years have seen the development of commercial multiplex PCR- or microarray-based procedures that enable detection and identification of pathogens from the liquid growth medium of blood cultures flagged as positive within 1 to 4 h (e39– e44).

A randomized controlled study showed that this method can help to optimize the antimicrobial treatment of bloodstream infections; no advantage could be demonstrated, however, with regard to mortality or duration of hospital stay (13). The counterarguments are the high costs, the additional workload, the lack of clarity concerning selection criteria for the samples to be investigated, the limited number of pathogens that can be detected, and, above all, the inauguration of MALDI-TOF MS as a procedure suitable for routine application (14).

MS procedures

Direct pathogen identification by means of MALDI-TOF MS from positive blood culture bottles has to be preceded by sample preparation using procedures such as lysis–centrifugation followed by ethanol/formic acid extraction. This quick technique—available either as an in-house adaptation or a commercial method—has shown identification rates of up to 80% and more within circa 20 to 40 min for detection of bacterial pathogens (7, 15– 17). In the case of yeasts, direct identification was possible in 62.5% of the samples (18).

Conventional diagnosis of bloodstream infections.

• Inoculation of the patient’s blood into a liquid medium (blood culture bottles) for primary pathogen culture

• Subculture on solid growth media

• Identification and antimicrobial susceptibility testing

However, this technique has not achieved universal acceptance. The reasons for this lie in the relatively large increase in complexity, the increased reagent costs, the moderate identification rates, and its categorization as an add-on diagnostic test, because it does not replace subculture on solid growth media. This led to samples being processed in batches (e.g., twice daily or less), for organizational reasons, so that the anticipated benefit of rapid results largely failed to materialize (10, 16, 19). A more economical procedure better suited to integration into daily routine thus had to be found (14, 20). Instead of the usual “mature” colonies after overnight incubation, subcultures from positive blood cultures were incubated very briefly on solid growth media and then used (7). The growth on freshly inoculated solid growth media was monitored closely. On the first visual determination of microbial biomass (a fine “haze”), MALDI-TOF MS analysis was carried out (figure 1). It emerged that microbial colony material can be used for MALDI-TOF MS as soon as the first faint growth becomes visible on solid media. The mean incubation time to successful species identification for Gram-negative rods was only 2.0 h; for Gram-positive cocci it was 5.9 h (3.1 h with preceding protein extraction). The identification results were always the same as for cultures incubated for 24 h (7). The procedure can readily be integrated into the routine of a microbiology lab and requires neither additional materials nor more time to perform. Subsequent studies confirmed the benefits (21– 24). Furthermore, this identification technique can be advantageously combined with rapid susceptibility testing (25).

Figure 1.

Enterobacter cloacae cultures incubated for various lengths of time on a solid growth medium (blood agar plates) and their use in diagnosis

Direct DNA-based pathogen identification from whole blood

Direct pathogen identification.

Direct pathogen identification by means of MALDI-TOF MS from positive blood culture bottles has to be preceded by sample preparation using procedures such as lysis–centrifugation followed by ethanol/formic acid extraction.

Various experimental—later also commercial—species-specific and broad-spectrum NAT procedures were introduced to greatly shorten the time required for microbiological diagnosis of bloodstream infections directly from whole blood samples (26, e45– e47). Depending on the test system used, multiplex PCR can produce a result after circa 3.5 to 8 h. Broad-spectrum NAT takes much longer (roughly an extra working day) owing to the necessary sequencing of the PCR amplicon. In contrast to culture-based procedures, NAT can be positive as soon as microbial DNA is circulating in the bloodstream (“DNAemia”) and also in the case of microorganisms that are non-viable or whose status is viable but nonculturable (VBNC). While this may be disadvantageous in the event of contamination, it can lead to improved sensitivity in the presence of pathogens previously treated with antibiotics.

Haze-like growth.

Microbial biomass can be used for MALDI-TOF MS as soon as the first faint (haze-like) growth becomes visible on solid media.

Depending on their spectrum, commercially available tests for direct detection of pathogens in whole blood are variably effective for detection of bacteria and yeasts responsible for bloodstream infections as well as, in some cases, marker genes for some multidrug-resistant pathogens (26, 27, e48). These multiplex PCR systems for detection of specific pathogens are limited to the microorganisms that most frequently cause bloodstream infections (circa 25 to 90 pathogens, depending on the test system used) and yield false-negative results for rarer pathogens. Lowered analytical cutoff values to artificially decrease sensitivity have been introduced for some coagulase-negative staphylococci and nonhemolytic streptococci, which may be present as contaminants on the skin. These lowered amplification cycle number cutoffs may lead to under-reporting of infection-related bacteremia involving these pathogens in vulnerable groups (e.g., neutropenic patients and premature neonates) (28).

The broad-spectrum PCR procedures require additional time for analysis of the amplicon. They are not restricted to individual pathogens and also enable identification of fastidious or non-cultivable pathogens, such as Tropheryma whipplei, bartonellae, mycobacteria, or molds. They are also indicated for use in situations where their swiftness in yielding a result, compared with culture-based techniques, is beneficial for treatment or prevention of infections.

Direct DNA-based pathogen identification from whole blood.

Depending on the test system used, multiplex PCR can produce a result after circa 3.5 to 8 h. Broad-spectrum NAT takes much longer owing to the necessary sequencing of the PCR amplicon.

Recent years have seen the publication of a large number of studies comparing the identification results of PCR and culture techniques (27, 29– 31, e49– e52). Early investigations into clinical benefit were limited to data from retrospective evaluations (32– 34, e50, e53) or from studies that were prospective but not randomized (35– 37). The first two randomized controlled trials examined the clinical benefit of direct multiplex PCR from whole blood (38, 39). In hematology patients with febrile neutropenia, the median time before the switch to targeted antimicrobial treatment was 21.4 h in the study group (multiplex PCR in addition to blood cultures) and 42.5 h in the control group (blood cultures only); p = 0.018 (38). A positive PCR result led to a change in treatment in 33% of cases. A study in critical care patients with sepsis showed that the additional use of multiplex PCR from whole blood achieved a significant reduction in the time from sampling to communication of the result compared with blood culture alone (15.9 h versus 38.1 h) (39). The mean time that elapsed between blood sampling and the switch from empirical to specific antimicrobial treatment was 18.8 h in the study group and 38.3 h in the control group.

Broad-spectrum PCR techniques.

These are not restricted to individual pathogens and also enable identification of fastidious or non-cultivable pathogens, such as Tropheryma whipplei, bartonellae, mycobacteria, or molds.

Modern NAT procedures are possessed of high diagnostic accuracy and can lead to faster detection of the pathogen and thus to earlier administration of specific treatment. If difficult or nonculturable pathogens are suspected, broad-spectrum NAT—as well as antibody-detecting techniques, not discussed here—can contribute to their identification. However, various limitations have prevented wide application of the NAT techniques beyond the study setting (box 2). The principal disadvantages are the absence of information on antibiotic resistance phenotype and the lacking potential for close characterization of isolates. Efforts are being made to counter the sensitivity problem of the NAT procedures, e.g., by selective concentration of microbial DNA (e54).

BOX 2. Limitations of direct detection of pathogens from whole blood.

• False-positive results owing to:

  • Detection of DNA from non-viable pathogens or from microorganisms that have already been adequately treated

  • Detection of DNA from contaminants (e.g., from transient bacteremia/fungemia or from a skin cylinder punched out in the course of sampling)

• False-negative results owing to:

  • Low sensitivity compared with culture (e.g., due to artificial adjustment of detection thresholds)

  • Interference with primers and probes by high amounts of nonmicrobial host DNA

  • Presence of inhibitors (e.g., heparin, iron, and immunoglobulins)

• Discrepancies with the result from conventional blood culture (no pathogen detected, different pathogen detected, simultaneous detection of two or more pathogens)

• Comparatively high costs (apparatus, kits, staff time)

• No or only limited (restricted to a few [known] resistance genes) information about antibiotic susceptibility

• No possibility of typing

• Currently limited research data available on most of the test systems

Antibiotic susceptibility testing and resistance gene detection

Although species identification in itself can yield information helpful for treatment decisions, it is the results of antimicrobial susceptibility that are crucial in guiding the final choice of targeted therapy. For this reason it is important to produce the susceptibility results as soon as possible in the presence of critical infections.

Conventionally, individual colonies from solid growth media are employed for susceptibility testing. In the diagnosis of bloodstream infections, this entails considerable delay owing to the incubation time needed after subcultivation from positive blood culture bottles.

Acceleration of susceptibility testing by changing the method of culturing

With the direct inoculation of test systems from positive blood culture bottles, one does not wait for the growth of subcultured colonies on solid growth media. Cell pellets produced by centrifugation of positive blood culture bottles are used for direct inoculation into susceptibility testing systems. For yeasts, the time to availability of results was shortened by 15.1 h (18). While the results for Gram-negative rods showed good agreement with conventional procedures, no satisfactory correspondence was achieved for Gram-positive cocci and yeasts (18, 40) (e55).

Modern NAT techniques.

Modern NAT procedures possess high diagnostic accuracy and can lead to faster detection of the pathogen and thus to earlier administration of specific treatment.

One elegant and cost-efficient strategy is based on using the biomass of briefly incubated subcultures. The subcultures used for susceptibility testing are incubated on solid growth media for only a very short time. The agreement with standard testing was found to be over 99 % (25). The mean incubation period for the subcultures was only 3.8 h for Gram-positive cocci and 2.4 h for Gram-negative rods. The time saved was thus approximately one day, without any added costs or complexity. This method can readily be combined with rapid identification by means of MALDI-TOF MS from subcultures incubated briefly on solid growth media (14).

Use of MALDI-TOF MS for accelerated detection of resistance and for susceptibility testing

The time savings and lower costs associated with MALDI-TOF MS have prompted various efforts to adapt this technique for the detection of individual resistance mechanisms and also, recently, as an alternative for the universal susceptibility testing. For instance, antibiotic-related alterations of the pathogen’s proteome may result in changes in mass that can be demonstrated by MALDI-TOF MS (e56). Another approach rests on changes in mass of the antibiotics used that can be caused by microbial resistance mechanisms. An example is hydrolysis of betalactam antibiotics by bacterial betalactamases (e57). These methods take around 0.5 to 4 h but are restricted to just a few resistance mechanisms and involve greater test complexity.

A new approach to the universal application of MALDI-TOF MS for rapid susceptibility rests on growth-based differentiation into “susceptible” and “resistant”—similar to conventional susceptibility testing. Antibiotics are added in breakpoint concentrations, and MALDI-TOF MS is used to find out whether microbial growth of the isolates being tested has taken place (=resistant) or not (=susceptible) (e58). This procedure was recently shortened and simplified still further by the design of a method involving incubation of microdroplets of the isolates directly on the MALDI-TOF MS targets (direct-on-target microdroplet growth assay, DOT-MGA) (e59). In a pilot study, this technique—still in development but potentially suitable for routine application—took 4 to 5 h to complete susceptibility testing for several antibiotics with sensivity and specificity similar to those of the current standard procedure (e59). Moreover, DOT-MGA offers the advantage of parallel identification and susceptibility testing in one assay and can also be employed for susceptibility testing directly from blood cultures (e59, e60).

Direct DNA-based detection of resistance genes in whole blood samples

Efficiency of direct susceptibility testing.

While the results for Gram-negative rods have shown good agreement with conventional procedures, no sufficiently good correspondence has been achieved for Gram-positive cocci and yeasts.

Ideally, in monoinfections, the microbial DNA isolated directly from samples of whole blood could be used not only for pathogen identification, but also to detect the genes or, if mutation related, the DNA segments determining the antibiotic resistance. However, the PCR multiplex systems in current use are stretched to their technical limits by the detection of the pathogens most frequently causing bloodstream infections, so that one either dispenses with molecular detection of resistance or includes only individual resistance genotypes. In the latter case, detection ensues either in parallel, in one assay, or, if the primary PCR points to a corresponding pathogen, in a second PCR. The commercial test systems available so far extend to the detection of the mecA/mecC genes as a sign of MRSA isolates or demonstration of the vanA-/vanB genes as a sign of vancomycin-resistant enterococci (VRE).

Pre-analytics, standardization and quality assurance

Pre-analytic factors (eSupplement 2) (12, 38, e61, e62) and aspects of standardization and quality assurance (eSupplement 2) exert considerable influence on the quality of the results of microbiological diagnosis of bloodstream infections (e63).

Summary

The modern microbiological diagnosis of bloodstream infections is moving close to the time of clinical suspicion of the presence of infection, but is still far from having any direct influence on the decision regarding initial antibiotic treatment. MALDI-TOF MS, combined with the use of briefly incubated solid growth media, has contributed to a massive shortening of the time needed for pathogen identification and susceptibility testing. PCR-based rapid tests are helpful in the event of unsuccessful culture of bloodstream infection pathogens, but cannot replace culture. The benefits of the new techniques can take effect only if pre-analytics, the workflow, the times required for each step, and quality management are all optimally adjusted and if the principal indications and limitations of the individual diagnostic procedures are known to the requesting physician.

Supplementary Material

eSUPPLEMENT 1

Basic principles of DNA-based and mass spectrometry methods

Broad-spectrum nucleic acid diagnostics

If a patient’s medical history and clinical presentation suggest no clear initial diagnosis and conventional diagnostic procedures (culture, serology) do not attribute the infection event to the presence of one or more pathogenic microorganisms, the treating physician can request so-called “broad-spectrum” nucleic acid-based diagnostics, consisting of DNA amplification followed by sequencing or probe identification of the amplicons (figure 2). These NAT detection methods, also referred to as “group-specific,” “broad-range,” “eu-/panbacterial,” or “panfungal,” rest on the presence of certain genes or gene segments which, in at least parts of their nucleic acid sequence, have been conserved by “evolutionary pressure” and therefore display a high degree of homology among various species.

These gene sequences represent a certain gold standard in the definition of bacterial, fungal, and parasite species and their phylogenetic classification, and their high diagnostic potential can also be utilized for broad-spectrum pathogen detection. Some examples of systematically investigated “marker genes” are hsp65 (65-kDa heat-shock protein), rpoB (B subunit of RNA polymerase), and the genes for the ribosomal subunits (bacteria: 16S rRNA, 23S rRNA genes; fungi: 18S rRNA, 28S rRNA genes) or interposed sequence regions known as internal transcribed spacer (ITS) regions.

The ribosomal genes and ITS regions represent ideal target sequences for molecular biological species identification of cultured isolates and also—under certain experimental conditions—for direct detection from the clinical sample (2). The amplicons are subjected to automated DNA sequencing, and the nucleic acid sequences determined are compared with databases, whereby the quality of the database contents is decisive for the interpretation of the sequencing findings (e17, 64). The results are communicated either as clear concordance with reference sequences from a given species or, as observed more frequently in practice for the investigation of clinical specimens, a list of several phylogenetically related species in ascending order of sequence homology. This enables clear species attributions in many cases, sometimes with the aid of existing culture results or molecular genetic analysis of further cross-species target sequences.

Many in-house techniques have been developed for broad-spectrum nucleic acid diagnostics; to date, commercial procedures are available for detection of pathogens responsible for sepsis.

If the sample for investigation consists of normally sterile material (e.g., blood, but also cerebrospinal fluid, joint fluid, or tissue), amplicon sequence analysis usually yields clear species identification even in the case of fastidious or non-cultivable pathogens and previously unknown pathogens. This type of analysis fails, however, in the presence of (the DNA of) various microorganisms in the same sample, in which case overlapping of the different species-specific sequence signals hampers interpretation. The indication for these specialized investigations therefore depends on certain preconditions and constellations (box 2) (3).

The greatest problem in the technical implementation of broad-spectrum NAT techniques in the microbiology lab is the potential for contamination from the ubiquitous ambient bacterial and fungal DNA (in the air and on surfaces) and the physiological microbiotic flora of patients and staff (e16). Even reaction vessels and reagents employed for DNA isolation and amplification can become contaminated to a varying degree with bacterial or fungal DNA during production or use (e18). With sensitive broad-spectrum test systems, for example, DNA from Pseudomonas spp. can be demonstrated in many batches of aqueous reagents approved for PCR (e19).

Until a complete palette of “DNA-free” sample vessels and other reaction components are commercially available, together with closed systems for contamination-free processing and analysis, accreditable and thus reliable performance of broad-spectrum NAT investigations will necessarily be restricted to a small number of ready-made test kits from certain manufacturers (e.g., those producing sepsis PCR kits) and to specialized laboratories that conduct elaborate batch testing for individual reaction components and possess the necessary lab space and equipment.

Therefore, before ordering broad-spectrum NAT investigations the attending physician must always weigh the relatively high demands on technology, quality management, special reagents, and working time (and thus costs) against the anticipated benefit for the individual patient.

Whole-genome sequencing and whole-metagenome sequencing

The technical accomplishment of whole-genome sequencing (WGS) and whole-metagenome sequencing (WMS) has been made possible by so-called next-generation sequencing (NGS), so that these methods are well on the way to becoming routine (4, 5, e20). Their advantage lies in the potential for rapid (hours to a few days) simultaneous identification, genotyping, and determination of genetically coded traits of virulence and resistance. Even parallel analysis of the host genome is possible if the sample contains human DNA.

These procedures, starting with the pyrosequencing technology in 2005, permit complete presentation of the genetic information for individual pathogens (by WGS) or complete microbiota (by WMS) in a relatively short time (e65). Current NGS platforms yield a throughput of 0.1 to 6000 Gb, depending on the device used (for comparison: 0.0003 Gb for the conventional, so-called Sanger sequencing [e66]). However, the current reading lengths of around 50 to 400 bp for “short-read” techniques are still much lower than for the first-generation sequencing methods (Sanger), which achieved reading lengths of 0.5 to 1 kb. Massive parallel sequencing following amplification results in a large number of small portions which, after assembly, provide multiple coverage or reading depth of the genome. New “third-generation” sequencing methods (single-molecule real-time sequencing techniques, biological nanopore-based techniques), which have single DNA molecules as target and thus require no amplification, are capable of achieving DNA fragments of = 20 kb (e67– e69). Depending on procedure, particularly for the long-read techniques, the error rates may be relatively high (10 bis >30%), especially in complex areas of the genome with many repeats or phage insertions. For this reason, sufficiently high coverage and efficient algorithms for assembly and error correction are necessary to pinpoint errors and close gaps (e70).

WGS techniques therefore represent the ultimate tool for genotyping of microbial isolates, because they are capable of detecting individual nucleotide exchanges between isolates. Using these techniques, the transmission pathways of pathogens can be elucidated in detail and their dissemination in a health facility, region, or beyond can practically be traced. In the area of hospital hygiene, prospective WGS can document infection pathways continuously and in real time, and achieves this at a lower cost than measures to prevent transmission (6, e21, e22).

The routine use of these techniques is currently limited particularly by the costs, the quality of databases, insufficient standardization, lack of IT capacity, and problems of “big-data” management.

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry

Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) enables marker-free analysis of large biomolecules by cocrystallization of matrix material and analyte (=sample) (e24). For diagnostic characterization, material from microbial pure cultures is applied to a plate together with the matrix and the resulting cocrystals are irradiated by a pulsed laser in full vacuum. The absorbed laser energy leads to ionization and release of ions that are then accelerated by an electrode in the electrical field generated, detected by an ion detector, and transformed into electrical signals. The time of flight of the pathogen-specific composite ions depends on their mass and charge. The measured values are compared with reference spectra, so the results are dependent of the size and quality of the databases used.

The culture-based areas of application for routine MALDI-TOF MS in diagnostic microbiology include especially identification of prokaryotic (bacteria, archaea) and eukaryotic pathogens (fungi, single-and multi-celled parasites) at species level, and sometimes at subspecies or even strain level (8, e25, e26, e71). Detection of specific viral proteins (e.g., norovirus capsid protein) by this means has also been reported (e72). Moreover, applications for detection of bacterial toxins, e.g., for Staphylococcus enterotoxins and Panton–Valentine leukocidin (PVL) have been described (e73, e74). Further reported uses are direct detection of microorganisms from positive blood cultures and from samples of urine and foodstuffs (18, e75– e78). An application of MALDI-TOF MS that is well on the way to routine use is identification of resistance in bacteria and fungi, or antibiotic susceptibility testing of clinical isolates (e56– e59, e79– e84).

eSUPPLEMENT 2

Pre-analytics, standardization, and quality assurance

Pre-analytic processes and factors

Particularly if add-on tests are used, costs can increase sharply, due to the test itself and/or increased personnel requirements. Not least for this reason, close attention should be paid to the necessary practical and administrative/organizational processes in the course of pre-analytics (e85).

These include determination of the indications, patient preparation, selection of the nature and number of samples for the add-on test systems, any diagnostic algorithms, transport and storage times (figure 2), and the organization of working times and diagnostic procedures in the laboratory. For example, excessively long transport and interim storage times or too small inoculates may lead to false-negative results in automated blood culture diagnostics (38, e61, e62).

Standardization and quality assurance

The quality requirements formulated at the beginning of the PCR era regarding standardization, reproducibility, and validity still hold true, not only for DNA-based methods: “Uncritical use of PCR procedures on unproven specimen types or the application of these methods under inappropriate conditions amplifies the potential for error and leads to diagnostic misinterpretations” (translated from the original German in [1]). Naturally the same requirements apply to MALDI-TOF MS-based techniques.

With the growing acceptance, widening use, and especially the ever-increasing variety of closed analysis systems, ready-prepared commercial test kits, and in-house protocols for nucleic acid detection of relevant pathogens, there is an increasing need for suitable external quality assurance measures. These should preferably take the shape of centrally organized inter-laboratory ring trials, but there could also be bilateral or multilateral exchange of well-characterized sample materials between laboratories.

While after successful accreditation and/or certification of a diagnostic facility only a well-organized workflow and reproducible, standardized implementation of precisely predefined protocols can be assured (lab-internal quality management), external quality control measures are aimed mainly at testing the sensitivity and specificity of NAT protocols and the assurance of precise and above all comparable results in diagnostic routine. Thus, internal quality management and external quality control are complementary.

Diagnostic institutions profit in more than one way from inter-laboratory ring trials or applying comparable quality assurance measures, as in this way they formally comply with the demands of certification and accreditation guidelines and also fulfill the requirements of the Richtlinie der Bundesärztekammer zur Qualitätssicherung laboratoriumsmedizinischer Untersuchungen (German Medical Association Guideline on Quality Assurance in Medical Laboratory Examinations; RiLiBÄK, part 3) (e63). Moreover, professional societies are also striving to achieve quality assurance and standardization of examination techniques, as exemplified by the microbiological/infectiological quality standards (MIQ) of the Deutsche Gesellschaft für Hygiene und Mikrobiologie (German Society for Hygiene and Microbiology; DGHM) (e86), e.g., MIQ 30, Qualitätsmanagement im medizinisch-mikrobiologischen Laboratorium (Quality Management in the Medical–Microbiological Laboratory); the norms of DIN series 58959, Qualitätsmanagement in der medizinischen Mikrobiologie (Quality Management in Medical Microbiology) (e87); and the Checkliste Mikrobiologie und Hygiene – Molekularbiologie in der Infektionsdiagnostik (Checklist Microbiology and Hygiene—Molecular Biology in the Diagnosis of Infections) in the Handbuch für die Akkreditierung Medizinischer Laboratorien (Handbook for the Accreditation of Medical Laboratories) (e88). All of these documents are published in German.

The Society for Promoting Quality Assurance in Medical Laboratories (Gesellschaft zur Förderung der Qualitätssicherung in medizinischen Laboratorien e.V.; INSTAND) has set up a German-language journal entitled GMS Zeitschrift zur Förderung der Qualitätssicherung in medizinischen Laboratorien (GMS Journal for Promoting Quality Assurance in Medical Laboratories) (www.egms.de/static/de /journals/index.htm) (e89). Here the participants in INSTAND ring trials receive reports of potential sources of error and further information on current issues in quality assurance.

Further information on CME.

• Participation in the CME certification program is possible only over the Internet: cme.aerzteblatt.de. This unit can be accessed until 3 March 2019. Submissions by letter, e-mail or fax cannot be considered.

  • The following CME units can still be accessed for credit:

  • „The Nomenclature, Definition and Distinction of Types of Shock“ (issue 45/2018) until 3 February 2019

  • ”The neurophysiology and treatment of motion sickness” (issue 41/2018) until 6 January 2019

  • ”The diagnosis and treatment of anxiety disorders” (issue 37/2018) until 9 December 2018

• This article has been certified by the North Rhine Academy for Continuing Medical Education. Participants in the CME program can manage their CME points with their “uniform CME number” (einheitliche Fortbildungsnummer, EFN).

• The EFN must be stated during registration on

• www.aerzteblatt.de (“Mein DÄ”) or else entered in “Meine Daten,” and the participant must agree to communication of the results. The 15-digit EFN is found on the CME card (8027XXXXXXXXXXX).

CME credit for this unit can be obtained via cme.aerzteblatt.de until 3 March 2019.

Only one answer is possible per question. Please select the answer that is most appropriate.

Question 1

What advantage does the additional microbiological diagnosis of bloodstream infections by means of nucleic acid amplification testing (NAT) have over culture-based procedures?

1. Consistently higher sensitivity

2. No need for preparation of the sample

3. Can be carried out by untrained staff

4. No quality assurance measures needed

5. Faster test result

Question 2

When is the use of broad-spectrum NAT indicated?

1. For detection of fastidious or non-cultivable pathogens and/or when its speed offers benefits for treatment or prophylaxis of infections

2. Only for experimental purposes in clinical trials

3. Exclusively for the detection of yeasts and molds

4. Exclusively for reasons of prevention (e.g., MRSA screening)

5. When mixed bacterial and viral infection is suspected

Question 3

What modification of the culture technique can accelerate susceptibility testing?

1. More glucose is added to the culture medium to improve the conditions for growth.

2. Positive blood culture bottles are centrifuged and the pellet is used directly for susceptibility testing.

3. The incubation temperature is raised to 39° so that the blood cultures reach the logarithmic growth phase sooner.

4. Sample material is first inoculated onto a solid growth medium, then individual colonies of bacteria are subjected to susceptibility testing.

5. Ampicillin is added to the blood cultures to select for resistant pathogens from the outset.

Question 4

What method is used for the microbiological diagnosis of bloodstream infections?

1. Exclusively NAT techniques are used.

2. It is primarily culture-based; the cultured pathogens are identified by various means and tested for antibiotic susceptibility.

3. Primarily serological detection methods are applied.

4. Cell culture techniques based on human host cells are employed.

5. Exclusively Gram-stains are used.

Question 5

What limits the speed of culture-based procedures in microbiological diagnosis?

1. The technical parameters of the incubators required

2. Suboptimal growth conditions due to standardized culture conditions with negative effects on the division rate

3. The generation time of the microorganism

4. The high amount of work involved

5. Exposure of the cultures to UV light and daylight spectra (including wavelengths of artificial light sources)

Question 6

What is the functional principle of diagnostic MALDI-TOF mass spectrometry (MS)?

1. Analysis of microbial nucleic acids (DNA) by MS

2. Analysis of microbial RNA by MS

3. Analysis of microbial, primarily ribosomal proteins by MS

4. Analysis of microbial islands of pathogenicity or other virulence-related elements by MS

5. Analysis of large biomolecules used for the typing of bacteria and fungi by MS

Question 7

What organisms are routinely diagnosed with MALDI-TOF MS?

1. Prion proteins

2. Bacteria and yeasts

3. Retroviruses

4. Double-stranded DNA viruses

5. Transfer RNA

Question 8

What is the conventional procedure for the diagnosis of bloodstream infections following blood sampling?

1. Inoculation into serum tubes, subculture on solid growth medium, antibiotic susceptibility testing on agar plates

2. Inoculation into blood culture bottles, subculture on solid growth medium, Gram staining for antibiotic susceptibility testing

3. Inoculation onto solid growth medium, Gram staining for pathogen identification, blood culture bottles for antibiotic susceptibility testing

4. Inoculation into the liquid medium of blood culture bottles, subculture on solid growth medium, identification and antibiotic susceptibility testing

5. Inoculation into blood culture bottles with various antibiotics, subculture and final antibiotic testing on solid growth medium

Question 9

From when can microbial colony material be used for MALDI-TOF-MS analysis?

1. After brief (30 min) incubation on solid growth medium

2. After full development of typical colony morphology

3. After appearance of at least three separate colonies at least 3 mm in diameter

4. After incubation for 1 h in liquid growth medium and brief (1 h) incubation on solid growth medium

5. After first visual detection of (haze-like) growth on solid growth medium

Question 10

What can lead to false-positive results when using methods for direct pathogen detection from whole blood?

1. Absence of means of typing

2. Presence of inhibitors such as heparin, iron, or immunoglobulins

3. Interference with primers and probes by high proportions of nonmicrobial host DNA

4. Low sensitivity compared with culture

5. Detection of DNA from non-viable pathogens or from microorganisms that have already been treated adequately

►Participation is possible only via the Internet: cme.aerzteblatt.de