Frequency of use of terminal points as an important risk factor for Legionella colonization in hospitals

Background: Hospitals have a different structure than domestic buildings. Legionella is a bacterium that can colonize the hospital water causing many outbreaks. It’s a Public Health problem and its prevention and control are essential to avoid nosocomial infections. The aims of this study are to evaluate several risk factors and parameters that can contribute to Legionella colonization in a hospital, thereby improving Legionella risk assessment process. Methods A total of 136 water samples (hot water and cold water) was investigated from different points of a hospital water network. These samples were tested for Legionella sp by three laboratories using different diagnostic tests: culture, polymerase chain reaction (PCR) and a method based on immunomagnetic separation (IMS). The results for these three techniques were combined and interpreted by three microbiologists with the aim of defining a new standard index. A multivariate analysis was performed by logistic regression to estimate the adjusted risk associated with the type of water, length of the pipe, chlorine, temperature, type of terminal point, period of the year, type of health care (ambulatory or hospital) and the frequency of use of the terminal point. Results: There was no statistically significant difference, on the basis of the new standard index, neither between laboratories nor between diagnostic tests. Positive results of this index were significantly correlated with the outpatient medical consultations, shower, frequency of use of the terminal points, and temperature. Logistic regression model revealed that outpatient medical consultations (p=0.058), shower (p=0.007) and the frequency of use of the terminal points (p=0.001) are predictors of Legionella colonization. Conclusions: The inclusion of rapid techniques (IMS, PCR) in the development of a new standard index offer increased sensitivity of Legionella detection improving classical

control strategies usually based on the culture method. The proportion of hospital watersystem sites testing positive, according to the defined standard index, was well correlated with the frequency of use of terminal points and the temperature, suggesting that the control of these factors may help prevent legionnaires' disease in the hospital.

Background
Legionellosis is an atypical pneumonia that is acquired when inhaling water drops contaminated with pathogenic bacteria belonging genus Legionella and they reach the lungs and replicate in macrophages (1). Legionellae are ubiquitous water-borne gramnegative bacilli. Legionellosis is primarily associated with two clinically distinct syndromes: Legionnaires' disease (LD), a potentially fatal form of pneumonia, and Pontiac fever, a self-limited, non-pneumonic illness, which usually resolves on its own without hospitalization or antibiotic treatment, within 2-5 days, so that it is generally not diagnosed. LD is the pneumonic form of legionellosis, with a case-fatality rate of 10% (25% for healthcare-associated infections). Mortality is higher in admitted patients and immunosuppressed patients, reaching 80% ever with proper treatment (2,3). In developed countries, the incidence has increased largely due to the introduction of clinical diagnostic tests, such as the test for detection of Legionella pneumophila serogroup 1 antigen in urine samples (2).
Legionellosis is a relevant pathology of increasing impact on human health. In 2016, in the 30 countries of the European Union, 6,560 cases were confirmed by the European Legionella Surveillance Network (European Legionnaires' Disease Surveillance Network -ELDSNet) (4). The exact incidence of legionellosis worldwide is difficult to measure and compare because countries have many differences in the methods of definition and notification of cases, having a low notification rate (2,5,6). Legionella is an important cause of both community-acquired and nosocomial pneumonia. The incidence of nosocomial infections is very low (7), but it represents a major public health concern mainly because nosocomial cases occur in a significant proportion in patients without any specific risk factor, and also due to the high lethality (50% in the Valencian Community, Spain). (8,9) The presence of bacteria in water is considered a prerequisite for infection (10,11). The growth of Legionella depends on several factors, among others, pH and hardness of the water, presence of old pipes, standing water, difficulty in maintaining the right temperature, biofilm formation, and the pipe length as an indirect factor (12) (13). In particular, health structures, including hospitals, health centers, residences, and dialysis centers, may present a risk environment for the transmission of LD due to the use of an old water network and the commissioning of medical devices with immunocompromised patients (14,15). Most of the sanitary water networks of the hospitals have been colonized by Legionella and, over time, it can cause the appearance of cases of nosocomial legionellosis.
Guidelines and regulations for the primary prevention of legionellosis (ie, Legionnaires' disease and Pontiac fever) have been developed by various public health and other government agencies. These guidelines are similar in recommending maintenance of building water systems; federal and other guidelines differ in the population/institutions targeted, the extent of technical detail, and support of monitoring water systems for levels of Legionella contamination. However, LD outbreaks are not uncommon despite guidelines and regulations are widely known and implemented (16). In Spain, for example, the regulation is focused on the levels of chlorine and temperature as risk factors, which is clearly insufficient. On this last point, the correct management of water quality, by including both cleaning procedures and rapid methods of analyzing Legionella in water, is considered as a critical issue to prevent LD outbreaks (17,18). This is due, among other factors, to the inability to predict the changes of the concentrations of free and intact Legionella in the aqueous phase, from which bioaerosols are generated (19).
All the factors explaining both colonization and infection of Legionella in sanitary environments are not yet clear (20,21). Volker et al. have tried to study some of these factors such as temperature, water stagnation and the pipe length. However, it must be considered that this study was not carried out on a sanitary structure, which involves specific difficulties and characteristics regarding a building for domestic use, so that the factors to be studied may be different. In addition, the author demonstrates that traditional culture-based methods are not always able to reveal the risk of Legionella contamination and water infection (13). For this reason, it is very important to study the factors that can be associated with an increase in the presence of Legionella in the water of a sanitary structure. Furthermore, a control of legionella colonization in a healthcare environment can improve health globally. In this context, the aim of this study is to evaluate the frequency of use of terminals points like a risk factor for Legionella colonization in a hospital (22). For qPCR trial (LAB3) each 1-liter water sample was mixed by shaking it and filtered through a 0.2 µm-pore diameter polycarbonate membrane filter (Millipore). The membrane filter was then removed and placed with 10 ml of sterile water and a portion of this concentrated was assayed by qPCR. This kit includes internal controls to control sample inhibitions or non-optimal reactions.

Data and sample collection
A new type of standard was used for diagnosis through consensus among three microbiologists from the three laboratories as it is described in the state-of-the-art, considering both culture, PCR and Legipid® Legionella Fast Detection test results (24).
Among the three participating laboratories cited above, a total of 527 water samples were examined to develop the standard index. Of these, 252 (47.82%) were cold water and 275 (52.18%) were hot water. Samples from areas where patients could not be exposed were excluded for the analysis, as in the case of water reservoirs and accumulators in the basement of the hospital. Once the index was defined, 136 water samples assayed by the LAB3 using all the analytical techniques (culture, PCR and IMS) was considered for this study. All the variables were recorded in an Excel spreadsheet.

Variables
The variables to be studied were the type of water, the pipe length, the chlorine, the temperature, the type of terminal point, the period of the year, the type of health assistance (outpatient or hospital), and the frequency of use of the terminal point. A terminal point was considered as frequent if this point was opened for at least 5 minutes every day.
The temperature factor was divided into three categories (<25 °C, 25-50 °C, and > 50°C), given that Legionella multiply where temperatures are between 20°C and 50º C and that the effectiveness of maintaining sanitary hot water at a minimum temperature of 55°C is significantly better than at 50°C for Legionella environmental control (25,26

Results
There was no statistically significant difference between the three laboratories (LAB1, LAB2 and LAB3) and there was also no difference between the IMS based method and the culture method (Table 1) confirming that they are equivalent methods as it is described in the validation of IMS-based method (27).
The Floors and Pipe length (distance from the terminal point) were not statistically significant factors. These variables are interrelated so that Pipe length was considered as adjustment variable provided that it is the most objective summary measure and it is cited in the literature. Although it is likely that there are others interacting factors.
The risk is associated with an increase in temperature above 25ºC and with hot water. The variable defined as temperature by ranges, which is best adapted to the theoretical behavior of Legionella growth, was included. The shower proved to be a terminal point of risk (OR = 2.05; p <0.05). Ambulatory health assistance was found to be a risk factor for colonization by Legionella (OR = 2.9, p <0.05).
In the logistic regression model the parameters of date, type of health assistance, terminal point, frequency of use, pipe length, temperature range and chlorine were considered. The May date (OR = 0.454; p <0.05) and Chlorine (OR = 0.032; p <0.05) turned out to be a protective factor. The odds-ratio (OR) for infrequent use was 12.187 (p <0.05). The increased risk is also confirmed for the shower point of sampleOR = 8.587; p <0.05).The ambulatory health assistance (OR = 12.248) was not, however, statistically significant (Table 3).
Usage frequency and temperature were taken into consideration. It was found that the hot water was the one that underwent the most changes. At terminal points of infrequent use (NFU) it had a greater percentage of temperatures less than 50 degrees. There was not any point of NFU showing a temperature higher than 55°C (which is present at points of frequent use -FU) (fig 1).

Discussion
Legionellosis was included in 1996 among notifiable diseases at the national level in Spain (Royal Decree 2210/95). In the Community of Valencia, the Basic Surveillance System, including legionellosis as a Notifiable Disease (OED), was developed by the Ordinance of March 4, 1997. Ten years later, in 2017, 236 cases of legionellosis were declared with an incidence rate of 4.7 cases per 100,000 inhabitants (9).
The national regulation currently in force is the Royal Decree 865/2003, which is an old text (28). Currently, this law notes the culture method described in ISO1173:1998 for the Legionella assay in water samples from cooling towers, but not explicitly for other risk installations. However, culture method is a slow method that can lead to a delay in taking timely steps on the facilities to prevent risk situations (23, 28,29).
Although recovery and isolation of Legionella colonies by culture is recognized as the gold standard method, well-known limitations of this method could compromise their utility in preventative or rapid control action. Among these drawbacks, we can highlight: (a) long time to confirm results (from 2 to 28 days are required), (b) changes on environmental water samples have been identified during their transport to laboratory, which might take up to 1-2 days, (c) likely presence of viable and infective but no-culturable cells, (d) poor sensitivity, and (e) a rate of inconclusive results up to 20% by interfering microbiota (30,31). According to the Royal Decree 865/2003, regulation on the prevention and control of legionellosis in Spain, a temperature between 25°C and 50°C is considered a risk factor for the proliferation and colonization of Legionella. In fact, thermal disinfection (hot water overheating) was the first method used to control Legionella in water distribution systems in hospitals (32,33). Water at 70 °C kills Legionella pneumophila in 10 minutes, while at 60 °C it takes 25 minutes (34). The temperature rise above 60 °C for several days ensures the success and absence of legionellosis (35). Its advantage lies in its minimum cost.
However, the loss of heat in some points of the hot water distribution network (heater, recirculation, representative points of use) may cause defective return valves in the faucets, providing generalized temperature losses due to the mixing of cold and hot water.
Moreover, although thermal disinfection is the most common, economical and effective treatment for the Legionella control, it does not eliminate the bacterial biofilm, which protects Legionella (36).
In our study, the average temperature is higher in the points of frequent use than those of infrequent use. The importance of maintaining temperatures above 50 °C, and even better above 55 °C (26), is well known.
Furthermore, more Legionella positive results were found using the new standard index than using a single technique, whether it is culture or IMS, only for temperatures around 40-50 °C (data not shown). For this reason, it is important to consider not-growth based diagnostic methods, able of measuring Legionella properties other than its growth in an artificial medium on a plate.
The importance of pipe lenght and water stagnation as risk factors for Legionella colonization have also been studied. These studies were conducted in residential facilities (13). The structural complexity of a health facility suggests means that pipe length, floor and pavilion cannot be used as risk factors, while it is important to consider both a correct and frequent use of terminal points (26) to prevent colonization and formation of biofilms.

Limitations
A limitation in this work is the difficulty to have a standard for the Legionella assay. The culture method is not a reliable test for the detection of Legionella in water with a preventative value. There is no consensus to establish an environmental diagnostic standard. The criterion used in this study increased the sensitivity of Legionella detection in water. However, it could be necessary to define an internationally validated standard in the future.

Conclusions
Legionnaires' disease (LD) is recognized as an important hospital-acquired disease.
Despite the several methods available, the optimal method to control hospital-acquired LD is not well established and their overall efficacy requires further evaluation. A combination of culture, PCR and IMS techniques offer the best index for detecting

Availability of data and material
Data are available from the authors upon reasonable request and with permission from the Hospital Clínico Universitario de Valencia.   Figure 1 Percentage of terminal points with a given temperature range (differently colored). Comparison between points of infrequent use (NFU) and points of frequent use (FU).