*2.1. Risk Management*

Using the checklist allowed us to verify that a water safety plan was in place in the hospital and a maintenance and control program was constantly applied to the water of the buildings and to the aeration systems. This should ensure a good quality of municipal water that can feed the DUWL when the switching is applied if the sterile water in the bottle runs out.

Despite the training activity for the correct adoption of the DU managemen<sup>t</sup> procedure (using only sterile reverse osmosis water, flushing between patients, self-contained water bottles disinfection, etc.), a low adherence to good practices was found in the hospital.

On the contrary, a good compliance with manufacturer's instructions for DU managemen<sup>t</sup> and use of biocides was observed in the private dental clinics, where the sta ff was not informed/formed neither on water risk assessment and managemen<sup>t</sup> nor on good practices guidance.

#### *2.2. Tap Water Results*

The microbiological quality of tap water varied between hospitals and smaller premises. *Legionella* spp. was not detected in three Private Dental Clinic (PDC) tap water samples (0/48), whilst in Hospital Odontostomatology Clinics (HOC), housing 13 dental units, *Legionella* spp. was repeatedly isolated in all tap water samples (30/104, 28.8%). The strains were identified as *L. pneumophila* sg 1 and sg 2–15, with a geometric mean of 4.91 ± 0.69 Log CFU/L and 4.38 ± 0.72 Log CFU/L, respectively. *Pseudomonas aeruginosa* and coliform bacteria were not isolated despite the fact that the total microbial counts at 22 ◦C and 37 ◦C were higher than the values recommended by Italian regulations (geometric mean of 2.79 ± 0.40 Log CFU/ml and 1.88 ± 0.41 Log CFU/ml, respectively). Free-living amoebae were recovered in 23% (3/13) of hospital DUs, but never in DUs housed in the PDCs. Among all cells microscopically positive to the culture examination, all PCR positive isolates belonged to *Vermoamoeba vermiformis* (identity of 99%). Despite the continuous application of the water safety plan, the microbiological quality of the municipal water remained low in the hospital and an inadequate concentration of residual chlorine-dioxide was detected (0.07 ± 0.14 mg/L). The mean temperature in cold water was 22.4 ± 1.6 ◦C, and the values were demonstrated to be related to *Legionella* concentration (R<sup>2</sup> = 0.51).

#### *2.3. Dental Unit Results before Shock Disinfection*

During the first sampling, water samples collected from hospital DUWLs showed a high prevalence of *Legionella*, which was detected in 31% (4/13) of dental units with a geometric mean 3.99 ± 0.61 Log CFU/L, often at di fferent sites on the device (Table 1). *Legionella* was isolated from inlets (3/4), spittoons (2/4) and from handpieces (4/4). Positive isolates were identified as *Legionella pneumophila* serogroup 2–15.

In PDCs, *Legionella pneumophila* serogroup 2–15 was isolated in 33% (2/6) of dental units during the first sampling, with a geometric mean of 4.15 ± 0.13Log CFU/L. *Legionella* was isolated from spittoons (2/2) and from handpieces (2/2).

All *Legionella* data collected before dental unit waterlines' disinfection are shown in Figure 1.

No coliform growth was detected in water samples. *P. aeruginosa* was isolated in high prevalence from handpieces and spittoons in 68% (13/19) of dental units. *P. aeruginosa* was detected in 33% (2/6) of collected samples, during the first sampling, from the spittoons of private clinics' dental units and in 85% (11/13) of dental units housed in the hospital clinic. *P. aeruginosa* was always associated with the presence of *Legionella*. In almost all water samples, the total microbial counts at 22 ◦C and 37 ◦C was ≥100 and 10 CFU/mL, respectively.

**Table 1.** Results of the microbiological analysis performed on water samples collected from dental units housed in a Private Dental Clinic and in the Hospital Dental Clinic before shock disinfection.


Note = TMC: Total Microbial Count; P/A: Presence or Absence; PDC: Private Dental Clinic; HDC: Hospital Dental Clinic.

**Figure 1.** Violin plot of *Legionella pneumophila* sg 2–15 (Log colony-forming units per milliliter (CFU/L)) detected from each sampling site of the dental units before disinfection.

*Brevundimonas vesicularis* was identified in two/six of the dental units housed in private clinics. Colonies were selected from Cetrimide Agar medium. All the percentages of positive dental units to waterborne pathogens are shown in Table 1.

Free-living protozoa (FLA) were detected in 46% (6/13) of HOC dental units. Among all FLA microscopically positive to the culture examination, one isolate showed band with approximate sizes of *Valkampfia.* All PCR positive isolates showed bands with approximate sizes of 800 bp (expected for *Vermoamoeba vermiformis*). The analyses unambiguously identified all samples as *Vermoamoeba vermiformis*, the sequences showing the highest identity (99%) with those accessible in GenBank. No water sample analyzed in this study was characterized for the presence of other FLA species, including *Naegleria* spp. Concerning *V. vermiformis* sequences, a phylogenetic analysis was also performed.

#### *2.4. Dental Unit Results after Shock Disinfection and Water Filtration*

In only one Private Dental Clinic was shock disinfection with hydrogen peroxide (HP) 3% v/v performed for pathogens (*P. aeruginosa* and *Legionella* spp.) in the DUWL, and filters were simultaneously installed at the inlet of each dental unit. Disinfection and filter installation showed a good e fficacy on *Legionella* spp., undetected after the treatment, as well as on *P. aeruginosa*. No other treatment was needed.

After 30 days from shock disinfection treatments with HP 3% v/v applied in 12 dental units in the HOC, total microbial counts at 22 ◦C and 37 ◦C resulted in the values being higher than the values recommended by Italian regulation (<100 CFU/mL and <10 CFU/mL, respectively) in almost all samples (67% and 92% respectively). *Legionella* spp. was detected in only one dental unit (8%), whereas *P. aeruginosa* was isolated in 83% of dental units (10/12), from handpieces and spittoons. The shock disinfection with HP 3% v/v showed a limited e ffect, with a recolonization period of about 4 weeks.

*Legionella* was eradicated after a shock disinfection with HP 6% v/v, applied after the installation of 31 days membrane filters at the inlet of each dental unit. The point-of-use water filtration showed good e fficacy in containing the entrance of *Legionella*, while HP disinfection had a good performance in controlling the growth of pathogens. After 30 days from disinfection with HP 6% v/v, *P. aeruginosa* was still found in 100% of DUWLs' samples (10/10), showing the higher persistence of this bacteria even after shock disinfection treatment. *P. aeruginosa* was not detected only after the third shock disinfection applied with a solution containing HP 4% v/v and surfactants.

All the results of the microbiological analysis performed on water samples collected from dental units housed in a Private Dental Clinic and in the Hospital Dental Clinic before shock disinfection are shown in Table 2.


