**2. Results**

## *2.1. Legionella Contamination*

The following results were obtained during the two phases of the study according to Italian Guidelines for prevention and control of legionellosis [7] that take into account the concentration of bacteria in relation to four levels of risk (<100, 101–100, 1001–10000, >10000 UFC/L) and the number of positive on the total number of samples collected. The measures to apply, in order to contain the risk, are di fferent if the percentage of positive samples are <20% or >20%.

#### *WTP1 Phase (October 2013 to March 2015)*

The data were obtained from the analysis of 53 hot water samples spread throughout the three buildings. We observed di fferent *Legionella* contamination trends in the MCH buildings (Table 1): 16/25 positive samples (64.0%) in Building 1, 13/23 positive samples (56.5%) in Building 2, and 3/5 positive samples (60.0%) in Building 3. The WTP1 phase was also compared with *Legionella* contamination data collected during the previous disinfection treatment involving the ClO2 mixture. Although we observed a change in the percentage of *Legionella*-positive samples in MCH (Building 1-2-3) from 95.0% to 60.0%, no statistical di fferences were observed in terms of *Legionella* contamination levels following the introduction of WTP 828. The analysis of *Legionella* contamination inside each building revealed a significant change in the mean *Legionella* levels only in Building 2 (*p* = 0.045) (Table 1).

#### *WTP2 Phase (September 2014 to October 2015)*

In the second phase of the study, we observed a reduction in terms of the percentage of *Legionella*-positive samples in MCH with respect to the WTP1 phase (from 60% to 35.8%) (Table 1). The same trend was also observed for the *Legionella* contamination levels (*p* = 0.0001). The results inside each building show a marked reduction in the percentage of *Legionella*-contaminated sites in Building 2 (from 56.5% to 7.0%) and Building 3 (from 60.0% to 34.0%); by contrast, the percentage of positive samples was only slightly reduced in Building 1 (from 64.0% to 58.1%).

The *Legionella* contamination levels displayed a significant di fference between the WTP1 and WTP2 phases for Building 2 (*p* = 0.046) and Building 3 (*p* = 0.048) (Table 1). No statistical di fference between phases was observed for Building 1, in which *Legionella* contamination levels of 1000 CFU/L were detected and, in accordance with Italian Guidelines [7], two shock treatments, increasing up to 50–60 mg/<sup>L</sup> of WTP 828, were performed (from February to March and from July to August 2015), resulting in a concentration of 25–30 mg/<sup>L</sup> at distal outlets.

The data collected from the WTP2 phase were also compared with *Legionella* contamination data obtained during disinfection with the ClO2 mixture. The comparison revealed significant di fferences for Buildings 2 (*p* = 0.0001) and 3 (*p* = 0.045) and no significant di fferences for Building 1 (Table 1).

During the study, the water reserves, softener, and tap water output sites were *Legionella*-free (below the detection limit of the culture technique used, i.e., 50 CFU/L).

To study the risk of *Legionella* disease that could be derived from the approach used during WTP 828 treatment (WTP1 and WTP2 phase) with the outcomes after ClO2 mixture treatment, we used two measures of risk provided by epidemiological studies: OR in a retrospective approach, and RR in a prospective study. Comparison of the ClO2 mixture with WTP1 (OR, 0.3) indicated that WTP 828 was not particularly e ffective in any of the MCH buildings (*p* = 0.048); however, comparison of the ClO2 mixture with WTP2 (OR, 15.44) revealed a significant improvement in *Legionella* control with respect to the latter (*p* = 0.0001). A prospective study indicated that the level of contamination during the WTP1 phase was higher than during the WTP2 phase (RR, 0.36, *p* = 0.002), showing a decrease in *Legionella* risk (Table 2).


**Table 1.** *Legionella* concentration in three buildings of Maria Cecilia Hospital (MCH) for each study phase.

\* Values are statistically significant at *p* < 0.05.

**Table 2.** Odds ratio and relative risk during the study phases.


\* Values are statistically significant at *p* < 0.05.

## *2.2. Legionella Typing*

The isolates from the WTP1 and WTP2 phases were serotyped and genotyped using standard techniques. The agglutination test permitted us to identify *L. pneumophila* serogroup 1 (SG1) and two *Legionella* species in 138/349 positive samples (39.0%). The SBT method assigned ST1 and ST104 to *L. pneumophila* SG1 isolates in 74/138 (53.6%) of the samples, while *mip* gene sequencing identified, *L. anisa* and *L. rubrilucens*, in 35/138 (25.3%); the remaining 29/138 (21.0%) samples contained a mixture of the previously described strains.

The results revealed that each HWN building was colonized by a different mixture of *Legionella* spp. Accordingly, Building 1 isolates were the most diverse with *L. pneumophila* SG1 (ST1 and ST104) and *L. species* (*L. anisa* and *L. rubrilucens*). All Building 2 isolates belonged to *L. pneumophila* SG1 (ST1 and ST104), and Building 3 samples demonstrated the presence of *L. pneumophila* SG1 (ST1), with some samples containing a single *L. species* strain (*L. anisa* or *L. rubrilucens*).

The serotyping and genotyping data about the bacterial concentration ranges (log10 CFU/L) are presented in Table 3.



96

During the study period, no significant association was found between *Legionella* colonization in the buildings and specific serogroups or strains. However, in Building 1, as after the two shocks treatment (February–March 2015 and July–August 2015), we observed a decrease in *L. pneumophila* SG1 levels, and detection of the other species, mainly *L. anisa* and *L. rubrilucens*. More experiments are still in progress.

#### *2.3. Pseudomonas aeruginosa and HPC Typing*

During the previous treatment by ClO2 mixture, the risk assessment plan for *Legionella* surveillance was performed without control of HPC at 36 ◦C and *P. aeruginosa*. After the introduction of WTP 828 treatment, during the study period (WTP1 and WTP2 phase), 349 hot water and 65 cold-water samples distributed among distal outlets, water reserves, softener, and tap water outputs of MCH were also analyzed for the presence of *P. aeruginosa* and HPC at 36 ◦C.

*P. aeruginosa* was not detected (as prescribed in D. Lgs 31/2001) [46] in either cold or hot water samples.

The HPC at 36 ◦C results for each building expressed as the mean concentration ± SD (log10 CFU/mL) were as follows: 0.82 ± 0.25 for Building 1 (0.48–1.20 log10 CFU/mL), 0.77 ± 0.65 for Building 2 (0.30–0.90 log10 CFU/mL), and 0.94 ± 0.35 (0.48–1.11 log10 CFU/mL) for Building 3.

At all sites, the contamination range was lower than the D. Lgs 31/2001 [46] limit of 20 CFU/mL (1.3 log10 CFU/mL).

#### *2.4. Physical and Chemical Parameters of Water*

The physical and chemical parameters linked to the quality of water after disinfection with WTP 828 were measured only during the WTP2 phase when relevant changes were made to the risk assessment plan. A total of 296 hot water and 65 cold water samples were analyzed. Physical and chemical data related to previous disinfection treatments and to the WTP1 phase are not reported because the *Legionella* surveillance during these phases took into account only bacteriological parameters.

The hardness, turbidity, and conductivity of the water (all of which are associated with the release of iron and total phosphorus) in the cold and hot water systems were not affected by WTP 828 treatment; these data are in agreemen<sup>t</sup> with recommendations established by Italian legislation [46]. In particular, the mean concentration of Ag+ remained lower than the detection limit (3 μg/L) and in line with WHO Guidelines for drinking water [33]. These results are shown in Table 4.



## *2.5. LD Surveillance*

During the study (WTP1 and WTP2 phases), 32 patients underwent urine antigen testing and other diagnostic tests because of suspected pulmonary signs of pneumonia. The negative results obtained confirmed the absence of cases of nosocomial legionellosis.
