**3. Discussion**

Legiolert is characterized by very easy and rapid sample preparation, with the additional advantages of avoiding the need for large sampling volumes, membrane filtration, treatments, plating, colony isolation and additional confirmation or identification. Furthermore, the Legiolert test reduces the time required to obtain confirmed results (seven days, rather than 10 or more days required by the plate culture method). In this study, potable water samples were analyzed by both plate culture and Legiolert methods and no significant differences were found when comparing results. Furthermore, the results of the analyses carried out using only 10 mL of water samples showed that the Legiolert test was equally reliable using 10 mL of water as using 100 mL.

Although the plate culture method is the gold standard for the detection and enumeration of *Legionella* in water samples, different laboratories may choose to follow different procedures, depending on the expected *Legionella* concentration in the samples they process, or even for economic reasons, affecting the reliability of the data when the same sample is analyzed by different laboratories. The plate culture method should be performed by accredited laboratories, according to norms recognized by the country's accreditation body. Among the methods, there are the ISO 11731:1998 method or the ISO 11731-2:2004 method (both of which have since been replaced by ISO 11731:2017); the 2007 American public health association (APHA) method; the Association Française de Normalisation (AFNOR) method NF T90-431:2018; or the U.S. Centers for Disease Control and Prevention (CDC) method [23–25]. For ISO 11731:2017 [15], depending on the matrix to be analyzed, the user may select from four methods, four treatments and four selective culture media, for a total of 14 possible procedural scenarios. Regardless of the method used, plate culture involves many steps and significant time requirements. *Legionella* monitoring, as part of the risk assessment analysis, concerns many different buildings such as hospitals, hotels, public offices and, in the near future, according to the revision of the European directive concerning potable water requirements, every potable water system. In this study, although we analyzed potable water samples according to ISO 11731:1998, the differences between this method and ISO 11731:2017 did not affect the recovery of *Legionella pneumophila*, since the

plating of the samples on BCYE medium, as suggested by ISO 11731:2017, would only improve the recovery of *Legionella* non-*pneumophila* species, which were excluded from this investigation. The MWY medium, although not included in ISO 11731:1998, is suggested by ISO11731:2017, and was adopted for this study because it has been known for many years to be the best for the recovery of *Legionella* from drinking water samples [26].

The newly drafted European Drinking Water Directive seems to have taken into consideration the emergence of many of these newer methods in Annex III part A, leaving to the national bodies the opportunity to choose the methods they find most appropriate for the purposes they specify [27]. In addition, the Legiolert method has recently been NF (Norme Francaise)validated by AFNOR certification and also included in the UK's Blue Book of validated test methods [28]. For many laboratories, the inclusion of testing in the Drinking Water Directive might lead to a large amount of work, time and financial expense. The Legiolert method may positively a ffect some of these di fficulties, as well as those linked to the managemen<sup>t</sup> of large volumes of water samples required for analysis by plate culture methods.

This study represents a confirmation of the reliability of the Legiolert method compared with the plate culture method, supporting conclusions from previous studies that documented the consistency of Legiolert for potable and non-potable water samples, analyzed according to ISO 11731:2004, which employs the filtration of 100 mL of water and the acid-treatment of filters which are directly placed on selective agar plates [16,29,30].

One limitation of the Legiolert method is that it is designed to detect only *Legionella pneumophila*, whereas other species remain undetectable. *Legionella pneumophila* is the most common species responsible for LD cases in Europe and, for this reason, in a few regions, such as France, Belgium, and the province of Quebec, Canada, it was decided to monitor only *Legionella pneumophila*, whereas in other countries, there is still a grea<sup>t</sup> debate on this matter.

Three fundamental factors can be identified in favor of monitoring exclusively for *Legionella pneumophila*. The first is risk—*L. pneumophila* is the species almost always cited in clinical cases and outbreaks and is the species most commonly found in the environment; the second is that laboratories may save time, human resources and money, and they can employ those saved resources to analyzing additional samples or locations, instead of identifying other *Legionella* species, which represent a much lower health risk; the third is that routinely monitoring only for the most pathogenic species of a bacteria is already an established practice. For example, *Pseudomonas aeruginosa* is routinely monitored, rather than all species of *Pseudomonas*.

At the same time, it is well known that other *Legionella* species are pathogenic to humans, although they represent a fraction of infections, with the exception of *Legionella longbeachae*, which is found in soil, rather than water, and is mostly detected in Australia and New Zealand. However *Legionella longbeachae* is beginning to be isolated also in EU/EEA, representing the 2.5% of isolated species in 2018 while other known and unknown species of *Legionella* were detected only in 3.3% of notified cases (European Legionnaires' disease Surveillance Network annual meeting 2019, unpublished data). It should be noted that the identification of species other than *L. pneumophila* su ffers from extensive use of the urinary antigen, which exclusively detects *Legionella pneumophila* serogroup 1, and from the medium used for the isolation of *Legionella*, which has historically been optimized for *Legionella pneumophila*. Therefore, many cases caused by other species might not be detected even by culture for this reason. Until a suitable medium for growing other *Legionella* species is developed, a routine PCR test in diagnosing human specimens, capable of distinguishing between *Legionella pneumophila* and other species, should be adopted in order to identify the real burden of Legionnaires' disease, as already demonstrated in a few countries [31–37]. The results of these studies will be able to confirm the real incidence of infections caused by other *Legionella* species and consequently to address the choices on what should be the focus of monitoring in the environment.

The imminent introduction of the new drinking water legislation concerning the monitoring of an increasing number of water systems, however, will probably lead to streamlined choices aimed

at reducing health risk by researching the most pathogenic and prevalent species present in the environment. Despite this, it must still be considered that for specific countries where other species of *Legionella*, such as *Legionella longbeachae*, are the prevalent in specific non-water matrices such as compost, and are an increasing cause of LD cases, Legiolert should not be utilized [1,31,35,37].

Concerning the enumeration of *Legionella pneumophila*, in this study, the most probable number did not provide any count in four of the 123 water samples, which instead tested positive by plate culture, though they were at lower concentrations of 50, 150, 300 and 650 CFU/L, respectively. For any Legiolert test, the limit of detection is 1 MPN, independent of the analyzed volume—a limit low enough to theoretically match that of plate culture. During this study, we used 100 CFU/L as the limit of detection for plate culture. We therefore suppose that the four samples which were negative in Legiolert but positive in plate culture could potentially be due to experimental errors. They could be, for example, faint colors of the wells not recognized as positive by the users. Unfortunately, these samples or isolates could not be tested again, as *Legionella* colonies were not kept for further investigations for detectability using the Legiolert test. The same bias can be considered for the eight TNTC samples tested by the 100 mL Legiolert protocol, which could have been included in the analysis if the original sample had been diluted or run with the 10 mL Legiolert protocol. No samples were found positive by Legiolert and negative by plate culture.

Although the data obtained showed that the two methods were comparable, a higher number of water samples at low *Legionella pneumophila* concentrations should perhaps be analyzed in order to assess any possible limitations with the Legiolert test.

In conclusion, Legiolert may be considered a valuable test for the detection and enumeration of *Legionella pneumophila* in potable water samples, and it can be used as a valid alternative to the traditional plate culture methods, especially considering the simplified protocol and the ability to employ smaller sample volumes to obtain the same quantification. Finally, it can be extremely useful when it is known that there is a prevalence of *Legionella pneumophila* in the water system under investigation.

#### **4. Material and Methods**

Over a 4-month period, 123 potable water samples were collected from hospitals, health care facilities for elderly people and industries located in 8 cities of the northern and central regions of Italy.

#### *4.1. Enumeration of Legionella pneumophila and Legionella Spp. by ISO 11731:1998*

Water samples were collected in 2-L bottles (1.5 L collected) and, after proper mixing, 1L of each sample was analyzed by the culture method according to ISO 11731:1998. Water samples were collected according to the protocol contained in the Italian guidelines for *Legionella* [38] and were stored at 5 ◦C ± 3 ◦C until they were delivered (within 24 h) to the Italian reference laboratory for *Legionella,* where all samples were analyzed.

The sample was filtered through 0.2-μm polycarbonate membranes and the membranes were transferred to 10 mL of the same sampled water and were solubilized by vortexing. From the concentrate, three MWY (Modified Wadowsky-Yee, Oxoid, Thermo Fisher Diagnostics Limited, Cheshire, UK) agar plates were inoculated by spread plating: one plate with 0.2 mL of the concentrated sample, one with 0.2 mL of the concentrated sample pre-treated with acid and one with 0.2 mL of the concentrated sample pre-treated by heating to 50 ◦C ± 1 ◦C for X 30 ± 2 min in a water bath. All the plates were then incubated at 36 ◦C ± 1 ◦C, with 2.5% CO2 for ten days. Presumptive *Legionella* colonies were confirmed by sub-culturing at least five colonies on BCYE agar plates with and without L-cysteine. The latex agglutination test (DR0800, Oxoid, Thermo Fisher Diagnostics Limited, Cheshire, UK) was utilized to obtain species information.

This study was carried out prior to the publication of the revision of the ISO 11731:2017 norm and the ISO 11731:1998 procedure was therefore applied. Regardless, the ISO 11731:1998 procedure adopted is still included in the new ISO 11731:2017 and is applicable for use with potable water samples, particularly when no information about the range of the *Legionella* concentration is known.

#### *4.2. Enumeration of L. pneumophila by Legiolert*/*Quanti-Tray*/*Legiolert*

The Legiolert test detects *Legionella pneumophila* through bacterial enzyme detection technology, which utilizes a substrate present in the Legiolert reagen<sup>t</sup> in a liquid culture to reveal the presence of *L. pneumophila*. Generally, 100 mL of the culture is analyzed and results are received in 7 days. Any turbidity and/or brown color greater than the negative control indicates positivity. Enumeration is based on MPN. In this study Legiolert was performed using both 100 mL and 10 mL of the original 1.5 L water sample collected. Each aliquot was processed and analyzed following the procedure outlined in the Legiolert instructions, using the Quanti-tray/Legiolert device. Quanti-trays were incubated for 7 days at 39 ◦C +/−0.5 ◦C in a humidified environment.
