bacteria are environmental organisms found in water, soil, and other environments. Typically, they are transmitted to humans via the inhalation of contaminated aerosols and they may cause fulminant pneumonia or mild Pontiac fever. In particular, nosocomial Legionella
spp. infection may affect immunosuppressed, transplant patients, and people undergoing aggressive chemotherapy [1
According to the European Center for Disease Prevention and Control (ECDC), the rate of Legionnaires’ disease in Europe in 2017 was 18 cases per million inhabitants. During the same period, in Europe, 9 out of the 30 countries reported 28 legionellosis outbreaks, either community or hospital acquired. A 30% increase in the number of reported cases in Europe was observed in 2017 compared with 2016, with an increase in the overall European notification rate from 1.4 to 1.8 per 100,000 population [3
The latest Italian data, related to 2017, give 2014 cases of Legionnaires’ disease in the Italian population, corresponding to a notification rate of 33.2 cases per million inhabitants. It must be noted that 2017 has seen an increase of 26% in the cases when compared with the previous three years. Of the 2014 notified cases, 124 (6.2%) were hospital acquired while 1580 (78.5%) were community-acquired infections [5
International and Italian guidelines for Legionnaires’ disease control suggest that large hospital water networks and high-volume hot water storage tanks provide optimal conditions for Legionella
colonization; the complexity of the water distribution systems and the presence of old and corroded pipelines and dead-end branches may promote the growth of Legionella
in hospital water networks [6
]. These critical points in water distribution systems cause a low water flow with a higher likelihood of biofilm development. Once microorganisms enter a system, they may be incorporated into biofilm, which is a primary cause of scarce water quality. Biofilm provides a nutrient source and protection from unfavorable environmental conditions. Several variables impact biofilm formation, including fluctuating water usage and transient stagnation [8
]. In a previous study, we reported the experience of an Italian hospital where a cost-effectiveness strategy for Legionella
control in the hot water distribution system was applied with continuous disinfection treatment and the installation of time flow taps (TFTs) in correspondence of all the dead-end branches, with the aim to increase the water flow of 196 L/day. Legionella pneumophila
sg2–14 was initially detected in all the sampling points despite the presence of chemical disinfection. After eight months of TFTs activity, Legionella
spp. was not detected anymore [12
Based on that preliminary experience, the aim of this study is the application of the same method in teaching hospitals hosting high-risk patients. In this manuscript, we report the data related to the decrease in Legionella spp. contamination in the hot water networks of four hospitals located in Northwestern Tuscany (Italy) from 2017 to 2019.
2. Materials and Methods
The healthcare setting, called Hospital 1, is a 60-bed Internal and Emergency Medicine Hospital. This structure has been active since 1950. The hospital architectural structure is a monoblock with a central plate on four levels. Hospital 2 (Oncology Unit) is a monoblock with a central plate with three levels which has been active since 1972. It has 30 beds. Hospital 3 is a 320-bed General and Transplant Surgery Unit which has been active since 2007. It is organized in five blocks with four levels. Hospital 4 is a 11-bed Senology Unit which has been active since 1987. It has a central plate on four levels.
2.2. Water Disinfection
Within the Water Safety Plan (WSP) implementation program, hot water system treatment and a systematic monitoring program with sampling at points of use began in all hospitals.
The hospital WSP is a self-control manual aimed to manage the microbiological water risk from waterborne pathogens, such as Legionella spp. This protocol is a technical guide which describes the control critical points of the water systems (dead-end branches, pipework materials, devices, etc.), the water samplings for microbiological tests, the disinfection procedure of water networks, and all responsibilities related to plant management.
Municipal drinking water was pre-filtered before entering the hospital hot water distribution system. Despite the old and galvanized steel-made pipelines, the hot water network was treated with continuous chlorine dioxide disinfection.
Continuous disinfection with chlorine dioxide was applied in water networks. This method was used one day before TFT installation.
In hospital water stations, a chemical feed pump, chlorine solution tank, and chlorine residual monitoring kit were installed.
Hot water was recirculated in order to obtain an energy saving. After the identification of all the dead-end branches present in the hot water distribution system, TFTs were installed corresponding to the most critical points.
They present an adjustable flowmeter in order to measure the scheduled water flow. TFTs were programmed in order to obtain a hot water flushing of 1 minute every 2 h (192 liters per day from each TFT), as described elsewhere [12
The TFTs, continuous chlorine dioxide disinfection, and municipal water pre-filtration plant were simultaneously installed in Hospital 1, Hospital 2, and Hospital 4. Four TFTs were placed in these buildings.
In Hospital 3, the continuous chlorination method was already present before TFT installation. Seven TFTs were simultaneously installed.
2.3. Hot Water Samplings and Legionella spp. Detection
According to all the hospital WSP programs, the hot water distribution system was sampled at the recirculation point of the central heating system (R), at the exit of the hot water production boiler (B), at each TFT, and at two points of use without TFTs (P1 and P2).
TFTs were installed close to some dead-end branches of the water networks (storage rooms, kitchens, and lockers rooms), while P1 and P2 were chosen from medical stays or surgery areas.
Legionella spp. detection and physical–chemical parameter assessment (total chlorine concentration, pH, and water temperature) were performed according to the following schedule:
One week before TFT installation (Time 0 or T0);
The day after TFT installation (Time 1 or T1);
Seven days after TFT installation (Time 2 or T2);
Fifteen days after TFT installation (Time 3 or T3);
One month after TFT installation, and then regularly on a monthly basis (Time 4 or T4, Time 5 or T5, etc.).
All physical–chemical parameter were measured in triplicate before and after TFT installation.
Samplings were performed after five minutes of water flushing. Hot water was collected in sterile bottles having sodium thiosulfate (20 mg/L). Temperature and total chlorine values were detected after the flushing.
Total chlorine concentration was determined by the colorimetric Visocolor HE® test (Macherey-Nagel, Düren, Germany).
Isolation of Legionella
spp. in hot water samples was performed as suggested by the international standards procedure ISO11731 [13
]. Briefly, one liter of water was filtrated through a 0.2 µm diameter membrane (Millipore, Billerica, MA, USA); the membrane was then immersed in 10 mL of the same water and sonicated for 5 min, allowing the detachment of the cells. The suspension was thermally inactivated by incubation at 50 °C for 30 min. Afterwards, 0.1 mL of the suspension was plated on Legionella
BMPA selective medium (Oxoid Ltd., Basingstoke, Hampshire, UK) and the plates were placed in one incubator at 37 °C for 7–10 days within jars with a modified atmosphere (2.5% CO2
) Finally, 10% of the Legionella
-suspected colonies that grew on the medium were subjected to confirmation test using BCYE medium and BCYE without cysteine medium (Oxoid Ltd., Basingstoke, Hampshire, UK). Confirmed colonies were subjected to species and serogroup identification analysis using a multipurpose latex agglutination test (Legionella
Latex Test, Oxoid Ltd., Basingstoke, Hampshire, UK).
Chemical samplings were taken at the same points chosen for microbiology tests. Chemical analyses were performed one week before TFT installation, six months after TFT installation, and then regularly on a six-monthly basis. Chemical parameters such as iron ions, zinc ions, and trihalomethanes (THMs) were determined as established by Council Directive 98/83/EC [14
2.4. Statistical Analysis
The Kolmogorov–Smirnov test was performed to verify normality of distributions. For each hospital, Kruskal–Wallis test and Dunn′s test were used to compare Legionella
spp. counts detected before and after TFT installations, as described elsewhere [12
]. Power tests were carried out to estimate the sample sizes. The 1-beta values of the significant variables were >0.8, demonstrating that sample sizes were acceptable. The statistical analysis was fulfilled using IBM SPSS software package version 17.0.1.
According to literature, Legionella
spp. control is needed in hospital settings, mostly in high-risk areas hosting immunosuppressed patients [15
]. International and Italian guidelines [6
] recommend the application of control measures to ensure good microbiological quality in hospital water networks. The choice of an adequate and continuous physical or chemical disinfection system may prevent the occurrence of waterborne pathogen colonization in hospital water plants [17
], but the complete eradication of Legionella
spp. seems mostly improbable if the colonization is high and the water network is old, large, and complex, with critical points [18
]. In spite of the age of the water plant and the presence of continuous chlorination in Hospital 3, we detected the presence of high Legionella pneumophila
concentration in hot water samples. This is due to the low hot water use in the hospital, which caused the persistence of Legionella
spp. bacteria in biofilm, as described in our preliminary study [12
]. In fact, the presence of dead-end branches in water networks allows the biofilm growth in points where the disinfectant cannot be effective against microorganisms [19
Despite the different number of hospital levels or beds, the amount of dead-end branches detected in hot water networks were the same, with the exception of Hospital 3. In fact, we installed four TFTs in Hospital 1, Hospital 2, and Hospital 4, while seven TFTs were installed in Hospital 3.
The choice of TFT installation points depends on the kind of subjects (patients, healthcare workers, etc.) present in those areas. We avoided TFT installation in stays areas, surgeries, and medical offices in order to limit the exposition to contaminated aerosol during TFT installation.
These devices were mostly placed in storage rooms, kitchens, and locker rooms.
The TFT installation occurred in correspondence with dead-end sections in all the hospitals and the implementation of continuous chemical chlorination in Hospital 1, Hospital 2, and Hospital 4, and the municipal water pre-filtration devices allowed Legionella pneumophila
decontamination during the period between T1 (24 h) and T4 (one month). As shown in Figure 5
, Legionella pneumophila
colonized all hospitals at T0 and T1. Contaminations persisted in Hospital 1 for one month and in Hospital 3 for seven days, despite the presence of total chlorine concentrations higher than 0.25 mg/L. After two months, Legionella pneumophila
was no longer detected. Moreover, in all hospitals, the increase of temperature values was obtained from T0 to T1. This is due to the maintenance works on the hot water networks performed during TFT installation. Considering the low hot water temperatures detected at T0 (≤40 °C), we increased this value, in order to achieve boiler water temperatures of at least of 50 °C. According to recent studies, the increase in hot water temperature, lack of water stagnation, and application of good hydraulic practices play a role in Legionella
spp. control in water networks [20
Our results show that water flushing eases the chlorine dioxide efficacy in water networks. A reduction of stagnant water reduces biofilm growth, which is considered as a form of microorganism resistant against chemical disinfectants, such as chlorine dioxide, applied in water networks [23
Literature data has shown the impact of unsteady flow on drinking water quality. Considering that stagnation promotes bacterial accumulation, special attention should be given to the biological quality of drinking water where consumption is variable. Interactions between hydraulics and biofilm structures are complex, and bacteria may be released from the biofilm into water after only 1 h of stagnation, allowing water quality degradation [22
In conclusion, our study continues the assessment of TFT device applications from our previous work [12
]. The aim of the TFTs was to increase the efficacy of the water disinfection system. In fact, TFT installation is considered a good cost-effective method which helps maintain disinfectant levels throughout the system [26
]. Although the periods of contamination decrease were long, this method could represent a valid choice for improving drinking water quality and disinfectant efficacy, reducing the critical section of water plants and biofilm maturation in dead-end branches.