Diagnostics and Clinical Disease Treatment

Diagnostics and Clinical Disease Treatment

Paul Christian Lück

Abstract

The methods currently available to diagnose Legionnaires' disease are culture, urinary antigen detection, direct fluorescent antibody testing, detection of nucleic acid and detection of specific antibodies in serum samples. The advantages and limitations of each method are discussed. Presently, none of the diagnostic tests available offers the desired quality with respect to sensitivity and specificity. Culture should be obligatory, especially when hospitalized patients with underlying diseases are investigated. A positive culture is the prerequisite of molecular epidemiological investigations. Urinary antigen detection is a valuable tool in the majority of community-acquired cases when L. pneumophila serogroup 1 (sg 1) is the causative agent. In cases of nosocomial disease, when Legionella pneumophila serogroups other than sg 1 are frequent, this assay has limitations. The detection of nucleic acid is very promising, but needs further validation. The detection of antibodies in a patient's serum is of little use in the acute phase of the illness.

Several molecular subtyping techniques are in use to subtype L. pneumophila strains in epidemiological investigations. Legionella pneumophila is genetically very heterogeneous thus allowing an individual fingerprint of each strain. However, the majority of clinical cases are caused by a limited number of clones that cause disease worldwide.

The therapy of for Legionnaires' disease requires drugs that can access and are active intracellularly. Currently, fluorochinolones and macrolides are the most active agents. read more ...

Introduction

An outbreak of severe pneumonia among the participants of the American Legion convention in Philadelphia in 1976 led to the description of Legionnaires' disease and the isolation of L. pneumophila as the causative agent (McDade et al., 1977; Chapter 1). Meanwhile it is well known that Legionella bacteria inhabit all kinds of freshwater reservoirs. The current list of validated names in the genus Legionella comprises 50 species with at least 73 serogroups (Euzéby, 2006). The most common species, L. pneumophila, contains strains that are classified as three subspecies (Table 2.1). Based on antigenic structures currently 16 serogroups are defined. Some L. pneumophila strains cannot be serotyped by using the established panel of monoclonal and polyclonal antibodies (Helbig et al., 2007). Hence, further serogroups might be present. Some of the non-pneumophila Legionella species, e.g. L. micdadei, L. dumoffii, L. bozemanae, L. longbeachae, have been isolated from both human and environmental sources; others have been isolated only from the environment (Table 2.1). Infections by L. longbeachae are often associated with potting soil mixtures, especially in some areas of Australia. It is generally accepted that all species may cause disease, especially in immunocompromised persons.

Approximately 75% of Legionella infections are caused by L. pneumophila sg 1, whereas 20-30% are caused by other serogroups of L. pneumophila, and 5-10% by different Legionella species (McNally et al., 2000; Helbig et al., 2002; Yu et al., 2002; Doleans et al., 2004) (Fig. 2.1). It is a well-known phenomenon that MAb 3-1-positive strains are more prevalent among clinical strains, especially in cases of community-acquired and travel-associated legionellosis. Therefore, it is believed that such strains exhibit a greater virulence. So far, this has not been defined on the molecular level. Recent studies suggest that a subset of clones, most of which react with MAb 3-1, are responsible for the majority of cases, whereas many Legionella strains of other serogroups are of lower virulence (Borchardt et al., 2006; Harrison et al., 2007). In contrast, the distribution of strains isolated from nosocomial cases reflects the distribution in the environment (Doleans et al., 2004; Fig. 2.1).

The distribution of Legionella species and serogroups has great influence on the sensitivity of the diagnostic methods and therefore on the reported prevalence of legionellosis. This is reflected in the current case definitions that are used in epidemiological investigations (see Chapter 3). read more ...

Prevalence of legionellosis among respiratory infections

Legionellosis occurs most frequently as sporadic cases, either nosocomial or community-acquired. Travel-associated cases were collected and reported from a European-wide surveillance system (Chapter 3).

Depending on local factors, the age of the patients and risk factors, the proportion of legionellosis among pneumonia cases has been reported to range from 1% - 22% (Blanquer et al., 1991; Fields et al., 2002; Fine et al., 2004; Sopena et al., 2005). In an ongoing study on the aetiology of community-acquired pneumonia, we found in 112 out of 2556 patients (4.2%) urinary antigen and/or Legionella DNA in respiratory or serum samples (Lück et al., 2006). If one accepts that these data represent the true incidence of Legionella pneumonia, then about 10,000 to 20,000 cases of Legionella community-acquired pneumonia occur in Germany each year. Based on these estimates the annually incidence in the United States would be 25,000 to
50,000 cases. read more ...

Laboratory diagnosis

Culture of Legionella spp.

Culture is still the 'gold standard' among the diagnostic methods for Legionella infections. The medium necessary for the cultivation of legionellae is buffered charcoal yeast extract (BCYE) agar supplemented with antibiotics. However, some Legionella strains might be susceptible to the antibiotics in selective media. Therefore, antibiotic-free agar should be used as an additional culture medium for material with no or low content of oral flora (Stout et al., 2003). Since quality control of media designed for the recovery of Legionella spp. is difficult to achieve in routine laboratories, the use of ready-to-use plates is recommended (Lück et al., 2004). For quality control of in-house preparations, standard Legionella suspension are commercially available (e.g. Health Protection Agency,
London).

Since legionellae are environmental, aquatic organisms which do not colonize humans and cannot be isolated from healthy persons, the specificity of culture is estimated to be 100% (Stout et al., 2003). 'False-positive' culture results may occur if clinical samples are contaminated with water containing legionellae, although reports of such cases are very rare (Lightfoot, et al. 1991).

The sensitivity of culture for the diagnosis of Legionnaires' disease has been estimated to be in the range of 11-65% by retrospective studies usually performed in reference laboratories (Hayden et al., 2001; DenBoer et al., 2004; Lindsay et al., 2004). In a retrospective study conducted in our laboratory we found a sensitivity of 80%. In contrast, the percentage of positive cultures tends to be lower if studies are performed prospectively in routine laboratories (Blazquez et al., 2005; Lück et al., 2006). Bronchoalveolar lavage fluid, bronchial aspirates, lung biopsies, post-mortem tissue specimens and sputum are particularly suitable for culture, whereas pleural fluid is less suitable (Edelstein, 2000; Fields et al., 2002; DenBoer et al., 2004). Legionella has been cultured from extrapulmonary sites (Stout et al., 2003) only seldom, in most cases due to systemic spread of the Legionella bacteria. However, the frequency of extrapulmonary legionellosis has not been studied extensively.

Legionella colonies usually form within 3-5 days. Suspected colonies are subcultures on BCYE Agar and cystein-free BCYE agar or Columbia blood agar. Colonies that grow on the former but not on the latter are presumptive Legionella species. One of the limiting factors for cultivation of Legionella spp. seems to be the experience of the laboratory staff (Edelstein, 2000). In some instances, Legionella colonies have unusual morphology that might be not recognized (Stout et al., 2003) (Fig. 2.2). read more ...

Identification of Legionella species

The most important technique for the identification of Legionella in the clinical laboratory is the serological characterization of isolated strains. A fluorescein-conjugated monoclonal antibody (MAb) which recognizes an outer membrane protein of L. pneumophila is commercially available. This species-specific MAb detects all serogroups and can be used for rapid identification of L. pneumophila from clinical samples or environmental specimens (Helbig et al., 2007). As mentioned before, this species can be divided into at least 16 serogroups. The serogroup-specificity is based on the chemical composition of the lipopolysaccharide, and the division into serogroups is based on the reactivity with polyclonal antisera and monoclonal antibodies. A few serogroups, mainly sg 1 of L. pneumophila, can be divided into MAb subtypes. This differentiation is useful for epidemiological investigations (Joly et al., 1986, Helbig et al., 2002). Polyclonal antisera, either fluorescein conjugated or coupled to latex beads, are commercially available for many, but not all, Legionella species.

Of practical relevance is the differentiation of L. pneumophila sg 1 strains from all other serogroups of this species. For this purpose, latex-coupled antibodies can be obtained from different suppliers. In general, all polyclonal antisera can cross-react with a variety of other bacteria. Such cross-reactions present few problems for the identification of colonies isolated on artificial media because the results can be verified by their morphology and requirement for cysteine.

Recently an immunochromatographic assay that utilizes monoclonal antibodies was developed and evaluated. This method is suitable to confirm the majority of Legionella species isolated from clinical and environmental samples (Helbig et al., 2006).

Fluorescently labelled oligonucleotide probes can also used to verify that a given isolate belongs to the genus Legionella (Buchbinder et al., 2004). This test correctly identified 43 of the described species (C. Lück, unpublished).

By using polyclonal antisera, FISH or the immunochromatographic test, it is difficult or even impossible to identify isolated strains to the species level since many species and serogroups show considerable cross-reactions or cannot be distinguished by these assays. Therefore, the most accurate identification of Legionella isolates to the species level can be obtained by sequence analysis of the 16S rRNA genes (Rubin et al., 2005) or the macrophage infectivity potentiator (mip) gene (Ratcliffe et al., 1998). A web-based tool for the identification based on their mip gene sequence is available from the European Working Group on Legionella Infections (www.ewgli.org). Likewise, sequences from other genes, like gyrA, rpnB and rpoB, are also suitable for differentiation of Legionella species (Fedderson et al., 2000; Ko et al., 2002; Rubin et al., 2005). read more ...

Detection of Legionella antigen in urine

An antigen excreted with urine has been characterized as heat-stable, resistant to enzymatic cleavage, and of about 10 kDa molecular weight. These characteristics are typical for lipopolysaccharide components. L. pneumophila is now divided into at least 16 serogroups and several monoclonal subgroups (Helbig et a., 2007). Indeed, all assays for the detection of L. pneumophila urinary antigens show sufficient recognition of the antigens which are homologous to the serogroup/monoclonal subgroup used as immunogen for preparation of the antisera, i.e. L. pneumophila sg 1 (Dominguez et al., 1998; Helbig et al., 2001; Guerrero et al., 2004).

Presently, several ELISAs are commercially available (e. g. Binax, Biotest, Barthels). The specificities of all assays, which were mostly evaluated by testing urine specimens from patients with urinary tract infections or pneumonias caused by other pathogens, have been reported to be > 99.9% (Dominguez et al., 1998; Helbig et al., 2001; Helbig et al., 2003; DenBoer et al., 2004).

With respect to the sensitivity for L. pneumophila sg 1 infections, all assays are able to detect more than 92% of cases identified by culture (Helbig et al., 2001). When culture-proven cases caused by other serogroups and legionellosis diagnosed by serology were included, the reported sensitivities dropped to 75-90%. This decrease reflects that all assays available have sensitivities for non-serogroup 1 infections of less than 35% (Helbig et al., 2001). The Biotest ELISA seems to be slightly more sensitive in diagnosing legionellosis caused by serogroups 2 to 15 (Horn, 2001).

Recently, it has been reported by several investigators that the sensitivity of the urinary antigen correlates with the severity of illness (Yzerman et al., 2002; Blanques et al., 2005; Lück et al., 2006) and the presence of underlying diseases (Sopena et al., 2002). This seems to be plausible since the 'antigen load' is higher in severe cases and the processing of the antigen might depend on the immune state. The use of concentrated urine samples increases the sensitivity without decreasing the specificity (Dominguez et al., 1998; Yzerman et al., 2002; Guerrero et al., 2004). Since a concentration step is time and labour intensive, we resort to this approach when the urine assay gives borderline results (Horn, 2001).

Besides these, a rapid immunochromatographic assay (BinaxNow) has been on the market for several years. Using this assay it is possible to detect antigenuria within a very short time, and no laboratory equipment is required. The BinaxNow assay is slightly less sensitive than the ELISA formats (Guerrero et al., 2004). Recently, other immunochromatographic assays were released to the market (SAS Legionella test, Rapid U Test, SD Bioline). Evaluation of these newer tests revealed specificity values below 100% (Diederen et al., 2006a; Diederen et al., 2006b; C. Lück, unpublished). Keeping in mind the low prevalence of legionellosis (2-5%), a lower specificity might be problematic for a diagnostic assay. Furthermore, the sensitivities for all newer assays are lower than for the BinaxNow. Thus, at present, these newer tests must be used with great caution.

The advantages of urinary antigen detection are striking: specimens are easy to obtain and can be investigated repeatedly; antigenuria is detectable very early and, therefore, often provides the first evidence of Legionella infection. Furthermore, the test is very rapid, and it has a very high specificity. In addition, positive urinary antigen results can also be obtained in cases of non-pneumonic, Pontiac fever-like illness (Lück et al., 1992; Lüttichau et al., 1998; Burnsed et al., 2007). In most cases, the antigenuria ends after 10 to 14 days. Despite the initiation of appropriate treatment, antigenuria may persist for some weeks or months, but this persistence of antigenuria does not reflect a failure of treatment. It is significantly associated with immunosuppressive therapy (Sopena et al., 2002; Koide et al., 2004). Despite the great advantages of the urinary antigen detection, one should keep in mind that a negative urinary antigen result never excludes a Legionella infection. read more ...

Detection of Legionella by direct fluorescent antibody testing

Direct fluorescent antibody (DFA) testing of respiratory specimens is a rapid method for the detection of Legionella antigen in respiratory samples. A monoclonal, fluorescein-conjugated antibody against an outer membrane protein of L. pneumophila is commercially available (Fig. 2.3). This reagent is highly specific but gives occasional false-positive results with Staphylococcus aureus, probably due to non-specific binding to protein A. The test detects only L. pneumophila. Since 90-95% of all cases of legionellosis are caused by this species, this limitation seems acceptable. The sensitivity of DFA testing has been reported to range from 27% to 70% (Hayden et al., 2001; Lindsay et al., 2004, C. Lück, unpublished). However, sensitivity depends mainly on the type of specimen used, the technical equipment (e.g. a cytospin centrifuge for concentration of diluted specimens), and the experience of the laboratory staff. In contrast to the urinary antigen detection which is mainly sg 1-specific and therefore has a high sensitivity for community-acquired cases, the sensitivity of the DFA test is not different for cases of community-acquired, travel-associated or nosocomial legionellosis (C. Lück, unpublished). The protein antigen detected by this assay is not degraded after fixation with formalin; hence, the test allows the detection of the aetiological agent in formalin-fixed lung tissue, which is not possible with the other methods available. An important issue is the prevention of false positive results of DFA testing due to contact of clinical specimens with water containing Legionella, due to contaminated buffers or legionellae that have washed off of positive control slides.

For the non-pneumophila species that are rare causes of legionellosis, no equivalent test exists. The use of polyclonal antisera for detection in clinical samples is limited by cross-reactions to other Gram-negative bacteria. read more ...

Detection of Legionella by fluorescence in situ hybridization (FISH)

Fluorescently labelled oligonucleotide probes can also used to detect Legionellae in clinical samples (Hayden et al., 2001; Buchbinder et al., 2004). However, FISH is technically demanding and requires an experienced laboratory staff.

Detection of Legionella nucleic acids in clinical samples

Over the last years, detection of nucleic acid has been more frequently used to identify Legionella in clinical samples. Depending on the primers used, the polymerase chain reaction (PCR) assays detect either L. pneumophila, or several, or all, of the published species of the genus Legionella (DenBoer et al., 2004).

Recently, extensive technical advances have been made for the direct (real-time) monitoring of fragments generated by PCR. Since Legionella is not considered to be part of the normal human flora, the presence or absence of Legionella DNA in specimens is the main clinical criterion, rather than the quantity of the pathogen. Therefore, the advantages and limitations of real-time PCR for quantification purposes are not the subject of discussion in this article. Beyond quantification, several advantages make real-time PCR superior to conventional PCR, even for non-quantitative applications. First, the post PCR processing steps like gel electrophoresis, hybridization or sequencing of PCR products, will become obsolete since PCR products can be detected immediately and specifically. This results in a significant acceleration of the diagnostic procedure and an improved specificity of the reaction. In addition, real-time PCR abolishes the risk of contamination that occurs during processing of the PCR products.

Detection of Legionella nucleic acids in respiratory samples

The first application of PCR for respiratory specimens was reported in 1992 by Jaulhac et al. (1992). By investigating the performance of PCR in bronchoalveolar lavage (BAL) fluid specimens, the investigators established the principle that PCR is suitable for the detection of Legionella DNA in clinical samples.

The currently available data allow a preliminary estimation of sensitivity and specificity values for PCR relative to other methods. The high sensitivity for the detection of Legionella DNA in respiratory samples demonstrated by several studies suggests that PCR may exceed culture in its ability to detect Legionella in the above mentioned specimens (Cloud, et al. 2000; Hayden et al., 2001; Rantakokko-Jalava et al., 2001; Reischl et al., 2003; DenBoer et al., 2004; Koide et al., 2004). For respiratory specimens an external quality assurance (EQA) system has been operating in Germany since 2004 (www.instand.de). This EQA scheme is open to all laboratories and should be use as a part of laboratory quality assurance systems. The results of this EQA scheme revealed certain problems in terms of sensitivity and specificity that occur in routine laboratories, but these are likely to be overcome in the future. read more ...

Detection of Legionella nucleic acids in urine

The excretion of DNA fragments in the urine has been described for several bacterial pathogens suggesting the suitability of urine PCR for the detection of Legionella DNA. Maiwald and colleagues (Maiwald et al., 1995) demonstrated the practicability of PCR from urine samples and revealed positive results not only in 29 of 37 samples from infected guinea pigs but also in samples from 11 patients with pneumonia, nine of whom were tested positive by other methods (Maiwald et al., 1995). Whereas the sensitivity of PCR seems not to be superior to the detection of Legionella antigens in urine, the possibility of using primer systems with specificity broader than the serogroup level appears to be advantageous for diagnostic purposes (Helbig et al., 1999). In general, the reported sensitivity varies from 17% to 72% (Murdoch et al., 1996; Helbig et al., 1999, Matsiota-Bernard et al., 2000, Lück et al., 2006). Reasons for this remarkable variation are not yet known.

Detection of Legionella nucleic acid in serum samples

The detection of Legionella DNA in serum has also been described (Lindsay et al., 1994; Murdoch et al., 1996; Matsiota-Bernard et al., 2000; Diederen et al., 2007; Lück et al., 2006). Legionella DNA is detectable in all acute phase and convalescent phase sera from patients with confirmed legionellosis, but not in the sera from over 100 patients without evidence of legionellosis. This approach appears rather promising since collection of this kind of clinical samples is inexpensive and simple to perform. As for urine samples, the sensitivity of the DNA detection assays in sera varies remarkably from 29% to 82% (Lindsay et al., 1994; Murdoch et al., 1996; Matsiota-Bernard et al., 2000; Diederen et al., 2007; Lück et al., 2006). No DNA was detected in serum samples from patients with pneumonia due to other organisms.

Antibody detection in human sera

Among the methods for antibody detection, only the indirect fluorescent antibody (IFA) test has been evaluated and standardized (Stout et al., 2003). It is the test format distributed by most of the commercial suppliers. ELISA tests are also available from different suppliers but are of lower specificities (Boshuizen et al., 2003; Malan et al., 2003; Rojas et al., 2005). This might be acceptable in epidemic situations but needs to be used with great caution when single cases are investigated.

The sensitivity of serology is generally limited by the time required to develop a detectable antibody response during the course of the infection. Approximately 25-40% of patients seroconvert within 1 week of the onset of symptoms; about 10-15% of patients seroconvert as late as 6-9 weeks after onset of the illness; and approximately 20-30% of patients do not develop significantly elevated antibody titers, even after prolonged observation. This variation limits the overall sensitivity of serology to 70-80% (Stout et al., 2003). On the other hand, the specificity of seroconversion (fourfold titre rise) using L. pneumophila sg 1 antigen in the IFA test has been reported to be approximately 99%. In general, it is recommended that Legionella serology be performed with polyvalent conjugates detecting IgG, IgM and IgA antibodies, since the immune response in the course of Legionella infections varies with respect to different immunoglobulin classes. IgM-specific IFA tests have a limited specificity and are not in use.

It must be emphasized that the exact sensitivity and specificity of serology has only been thoroughly evaluated for IFA tests using L. pneumophila sg 1 antigen. Sensitivity and specificity of antibody detection by using other serogroups of L. pneumophila and of non-pneumophila legionellae are not precisely defined. Probably, the values are considerably lower than those for serogroup 1. However, the detection of antibodies can be used to identify infections by novel species if appropriate serum samples are available (McNally et al., 2000). Infections with other serogroups of L. pneumophila can be detected with serogroup 1 antigen in some, but not all, cases. Therefore, many laboratories perform Legionella serology with a number of antigens. Pools containing different antigens are commonly being used for screening purposes, but positive reactions must be confirmed by monovalent antigens, as non-specific reactions occur frequently with pool antigens.

The background frequency of elevated anti-Legionella titres in the normal population has been found to vary from 1% to 36% (Stout et al., 2003). This limitation should be considered when single elevated titres are used as diagnostic criteria. read more ...

Epidemiological subtyping of isolated Legionella strains

So far, a positive culture is the only method that allows the comparison of patient and environmental Legionella strains necessary to confirm or exclude a given environmental reservoir as the source of the infection.

Epidemiological investigations of legionellosis are often complicated by the ubiquity of legionellae in nature. Because the incubation period varies from 2 to 10 days, the length of stay in a hotel, private accommodation, public building or hospital before onset of clinical signs does not establish with certainty where the infection has been acquired. Recognized sources of Legionnaires' disease confirmed in epidemiological investigations are: warm and cold water supplies (shower, taps) in private homes, hospitals, hotels, public buildings, cruise ships, cooling towers with 'water based cooling', whirlpools, thermal springs, moistener/respirators, decorative fountains, humidifiers for food display cabinets, car washers.

Often different species, serogroups, and monoclonal subtypes of Legionella are isolated from a given environmental source (Lück et al., 1998; Visca et al., 1999; Beyrer et al., 2006). In contrast to this, simultaneous infection with multiple Legionella strains seems to be a rare event (Horbach et al., 1988; Lück et al., 1998; Buchbinder et al., 2004). Such reports again underline the differences in virulence of different strains as illustrated in Table 2.2. In each case provided, the causative strain was in the minority among environmental isolates.

In some cases, no corresponding environmental isolate could be found, though all suspected water sources were investigated (Jonas et al., 2000). In such cases, it might be that the patient acquired the infectious strain from other environmental sources that were not investigated, e.g. outside the hotel or hospital, during overnight stays in the nearby private accommodation, etc. On the other hand, the causative strain might not have been isolated and subsequently typed because the source of infection was decontaminated or the bacterial population in the environment had changed and/or the causative strain had been overlooked.

All the above mentioned problems lead to discussions as to the number of colonies which should be typed after primary isolation and to the preferable typing method(s). At present, we serotype at least six colonies from environmental samples. Further, a combination of antigenic and genomic typing systems applied as a step-by-step procedure is recommended for the identification of Legionella strains that cause the infection (Table 2.3).

Monoclonal antibody typing

Subtyping of Legionella pneumophila strains by using monoclonal antibodies (MAb) was the first technique used in epidemiological studies and is now a well established method (Joly et al., 1986). As major advantages, it is technically simple and quick to perform. The reactivity patterns are stable, although changes have been observed due to point mutations and deletions in genes involved in the synthesis of lipopolysaccharide (LPS) (Zou et al., 1999; Lück et al., 2001; Bernander et al., 2003). In general, genetic variation is of minor relevance, since such events occur at a relative low frequency under natural conditions.

The major disadvantages of MAb typing are the time-consuming and expensive establishment, maintenance and quality assurance of the hybridoma cell lines producing MAbs. Furthermore, the antigenic diversity, or number of subtypes, is limited. Thus, the index of discrimination, an important characteristic which reflects the ability of a typing system to recognize different strains as different, is in the range of 0.8. Nevertheless, MAb subtyping is applicable to L. pneumophila sg 1, which can be divided into at least 12 subtypes (Joly et al., 1986; Helbig et al., 2002). When the MAb type of the clinical isolate does not match that of environmental strain(s), these reservoirs can with high probability is excluded as the source of the infection. At times, the use of subgroup-specific and cross-reacting antibodies has allowed the subtyping of strains belonging to serogroups 2-15 (Helbig et al., 2007).

All MAb subtypes are named according to a reference strain that shows this particular reactivity pattern. This kind of data can easily be exchanged between laboratories. In a multicentre evaluation of typing methods for the epidemiological typing of L. pneumophila serogroup 1, the epidemiological concordance was very close to 1 (Fry et al., 1999).

Macrorestriction analysis (MRA) by pulsed-field gel electrophoresis

Subtyping of Legionella pneumophila strains by MRA is an excellent tool for subtyping Legionella species and was considered to be the gold standard for more than 15 years (Lück et al., 1998; Fry et al., 1999; Jonas et al., 2000). Macrorestriction patterns can be analysed both visually and by computer-aided methods. The index of discrimination is greater than 0.95, the value which is required for a good subtyping system. However, the intra- and inter-laboratory standardization and exchange of data are difficult (Fry et al., 1999). Nowadays, sequences based typing (SBT) has replaced MRA. read more ...

Amplified fragment length polymorphism (AFLP) typing

This technique uses a combination of DNA restriction by asgn endonuclease and amplification by the polymerase chain reaction. It was the first typing system standardized within the European Working Group on Legionella Infection (EWGLI) (Fry et al., 1999, Jonas et al., 2000; Fry et al., 2002;). However, its index of discrimination was lower than for MRA, and the inter-laboratory exchange was impaired by the subjective methods of evaluation.

Sequence-based typing

Sequence based typing (SBT) is a variant of multilocus sequence typing that employs variations from multiple chromosomal locations, or genes. Currently, the European SBT panel includes six L. pneumophila genes: flaA, pilE, asd, mip, mompS, proA. Thus, an SBT type comprises a string of the individual allele numbers of each of these genes separated by commas. The major advantages of SBT are stability of the marker, good discriminatory power if appropriate loci are selected, and flexibility, since additional gene loci can be investigated if necessary. Data are readily exchanged among laboratories either as sequence data or as designated alleles. Furthermore, SBT reduces the need to transport live bacteria, since nucleotide sequence determination from PCR products can be achieved from killed-cell suspensions, purified DNA, or clinical material. While SBT is particularly suited to long-term and global epidemiology, as it identifies a variation which is accumulating slowly within a population, the data can also be used to investigate single cases or outbreaks (Gaia et al., 2005). Currently, the allocation of the allele formula can be done using the EWGLI website.

Antibiotic therapy

In vitro susceptibility of Legionella

In vitro susceptibility testing can be performed on artificial media to screen for active agents. It must be considered that components of the media (e.g. charcoal) might influence the in vitro susceptibility data. Macrolides, chinolones, ketolide, quinupristin/dalfopristin, doxycyclin, imipenem, rifampicin and tigecyclin are Legionella factors that are active in vitro (Stout et al., 1998; Nielsen et al., 2000a; Hammerschlag et al., 2001; Edelstein et al., 2003; Stout et al., 2005).

Resistance to all clinically relevant substances can be induced in laboratory experiments (Dowling et al., 1985; Nielsen et al., 2000b; Jonas et al., 2003). However, so far, there is no evidence that resistance occurs in clinical situations. Case reports describing clinical failure and/or prolonged clinical illness were never related to the development of resistance to antimicrobials (Rudin et al., 1984; Kurz et al., 1988; Tan et al., 2001; Gläser et al., 2005)

Intracellular activity of antimicrobial agents against Legionella

As legionellae are intracellular pathogens, L. pneumophila has been cultivated in vitro in a number of macrophage-like cell lines. In this way, the intracellular activity of antimicrobials can be assessed (Stout et al., 1998; Jonas et al., 2003; Stout et al., 2005). Generally, these results show a good correlation with animal experiments. In earlier studies it was shown that although bacterial growth is inhibited by erythromycin and rifampicin, it recurs when the drugs are removed from the cells (Edelstein, 1998). This is in contrast to newer fluoroquinolones and macrolides, which kill intracellular Legionella bacteria and do not permit bacterial regrowth. Azithromycin is the most active macrolide, and it has a much higher activity than erythromycin against intracellular L. pneumophila (Edelstein et al., 1991, Fitzgeorge et al., 1993; Jonas et al., 2003). read more ...

Clinical experience in the treatment of legionellosis

The clinical experience of azithromycin in the treatment of Legionnaires' disease is known to be safe and efficacious (Edelstein, 1995; Plouffe et al., 2003). The vast majority of patients who receive monotherapy with intravenous azithromycin for 2-7 days, followed by oral azithromycin, are cured.

The efficacy of levofloxacin was reported in a study analysing six clinical trials encompassing a total of 1997 patients with community-acquired pneumonia. More than 90% of mild-to-moderate and severe cases of Legionella infection were cured; no deaths were reported (Yu et al., 2004). Recently, three observational studies comparing levofloxacin vs. macrolides (not Azithromycin) in the treatment of legionnaires' disease have been published (Blázquez-Garrido et al., 2005; Mykietiuk et al., 2005; Sabria et al., 2005). There were no significant differences in clinical outcomes among the groups of patients with mild or moderate pneumonia; but, in patients with severe pneumonia, levofloxacin was slightly more effective. The combination of rifampicin and levofloxacin provides no additional benefit but does increase the rate of side effects.

In summary, as compared with older macrolides in the treatment of Legionnaires' disease, levofloxacin appears to be associated with better clinical outcomes, including a faster resolution of pneumonia symptoms, a more rapid achievement of clinical stability, and a shorter hospital stay. Thus, monotherapy with levofloxacin might be regarded as first-line antimicrobial for treatment of Legionnaires' disease. However, it must be underlined that the direct comparison of azithromycin and levofloxacin in the treatment of Legionella pneumonia has not been performed. Thus, it can be recommended to use fluoroquinolones or azithromycin, rather than older macrolides, for the treatment of Legionnaires' disease (Pedro-Botet et al., 2006).

Delay in the initiation of appropriate antibiotic therapy for Legionella pneumonia significantly increases mortality (Heath et al., 1996; Gacouin et al., 2002; Lettinga et al., 2002). It is therefore recommended that anti-Legionella agents be included early in the empiric therapy of severe community-acquired pneumonia. However, there are different view points regarding the first line antimicrobial therapy that is generally recommended for treating pneumonia (File et al., 2004).

Conclusions

Increasingly recognized as a major cause of sporadic and epidemic community-acquired and nosocomial pneumonia, Legionella represents an important public health problem.

Urinary antigen testing is a valuable diagnostic tool for the rapid detection of Legionnaires' disease. The early recognition of Legionella pneumonia by using urine antigen testing may have contributed to the decrease in mortality observed in the last years. Since this test detects mainly infections caused by L. pneumophila sg 1, it is especially valuable in outbreak situations. Detection of Legionella DNA in respiratory, urine and serum samples is now established and recognized as a rapid and reliable diagnostic assay; as such, it should be used more often to detect sporadic and epidemic cases.

The frequency of diagnosis of legionellosis by culture has decreased significantly in the last years, impairing epidemiological investigations. To reverse this problematic trend, culture of respiratory materials should be performed whenever possible, especially in cases of severe pneumonia.

The detection of antibodies in patient's serum is still a good laboratory test. Although of little use in the acute phase of the illness, it is especially valuable in epidemiological investigations.

Systematic molecular subtyping of all environmental and clinical isolates by using sequence based typing is an important tool in the investigation of Legionnaires' disease outbreaks and will help to identify the sources of the infection rapidly and reliably. By this approach, the occurrence of further cases can be prevented. Furthermore, such investigations will improve our understanding of the epidemiology of legionellosis, knowledge that is fundamental to any scientifically proven risk assessment.

Newer quinolones and macrolides have greater activity on intracellular legionellae in cell culture models and improved efficacy in animal models as compared with older macrolides. Monotherapy with a quinolone (i.e. levofloxacin) may be the current treatment of choice for Legionnaires' disease, according to results of recent observational studies. However, direct comparison of azithromycin and quinolones has not been performed in clinical studies. The addition of rifampicin to the treatment with newer quinolones or macrolides probably provides no benefit, but does increase the rate of side-effects. Although the combination of quinolones with azithromycin is supported by some in vitro data, the relevance of this observation for patients with severe Legionella pneumonia is not yet clear.

Acknowledgements

Own work was supported by the Federal Ministry of Education and Research of Germany, Network of Competence in Medicine CAPNETZ. I would like to thank Carolin Dix, Sigrid Gäbler, Jutta Paasche, Kerstin Seeliger, Ines Wolf for excellent technical assistance.

from Legionella: Molecular Microbiology
See also: Real-Time PCR in Microbiology: From Diagnosis to Characterization

Biology Blogs

Conferences

Journals

Open Access Biology Home | PCR | Microbiology Books | Molecular Biology Books | Molecular Biology Gateway | PCR Protocols | Real-Time PCR |