Impact of wastewater treatment plants on microbiological contamination for evaluating the risks of wastewater reuse

Wastewater reuse represents a promising alternative source of water supply considering the water scarcity related to climate change. However, if not adequately treated, wastewater represents a source of microbiological health risk. The purpose of this work was to investigate the role of wastewater treatment on microbiological contamination by evaluating the possible risks associated with wastewater effluent reuse, taking into account new EU legislation (2020/741) on minimum requirements for water reuse. E. coli that produce Shiga toxins (STEC) and thermotolerant Campylobacter were monitored using an enrichment step associated with specific PCR, while Salmonella spp. and Legionella were detected with both cultural and molecular methods (PCR and q-PCR, respectively). Culture method was also used for the enumeration of different microbial indicators. The bacteria detection was compared in different wastewater plants with membrane bioreactor (MBR) system or with disinfection step with chlorine dioxide (ClO2). Moreover a comparison between molecular and culture methods was discussed. The results obtained showed good abatement performance for WWTPs equipped with MBR. The high concentrations of E. coli (range between 0.88 and 5.21 Log MPN/100 mL) and contamination by Salmonella spp. in effluent disinfected with ClO2 (17% of samples) showed the need to control the quality of this effluent. In addition, despite the absence of Legionella spp. with the culture method required by EU regulation, high concentrations of Legionella spp. (range between 2 and 7 log GU/L) and the presence of Leg. pneumophila with qPCR (15% of samples) highlight the need to carry out further investigations for reuse associated with aerosol formation (e.g. spray irrigation in agriculture). The results obtained underline that the MBR technology can be suitable for wastewater reuse applications allowing to achieve the requirement proposed by the new European legislation. More attention should be given to wastewater reuse of effluents treated with ClO2. The use of the molecular methods for pathogens detection in wastewater could allow a more precautionary risks estimation associated with reuse. The overall results highlight that an evaluation of the effectiveness of the wastewater treatments is required for the prevention of a possible risk to public health.

in terms of quality and quantity, further exacerbating the situation of water scarcity that already characterises many world areas. Moreover, the growth of the world population will determine in the future the increase in the global demand for water to satisfy domestic, agricultural and industrial needs [2]; in addition, a greater number of urban settlements and human activities will lead to an increase in wastewater production. In this context, wastewater can become the main alternative source of water supply, providing a resource, the availability of which is not subjected to seasonality and does not require additional depletion of groundwater and surface water bodies. However, treated wastewater remains a minor source of water supply: at present, 0.59% of treated wastewater in the world and 2.4% in Europe are reused [3].
To ensure that wastewater reuse does not pose a threat to human health, the efficacy of wastewater treatment and the monitoring of effluent quality to prevent and limit the spread of diseases represent the main topics. Recently, European regulation (2020/741) on minimum requirements for water reuse has been adopted with the purpose of facilitating the safe reuse of treated urban wastewater for agricultural irrigation. This EU regulation proposes an approach based on minimum requirements in terms of reference indicators or pathogens and risk management when wastewater is reused for irrigation [4]. The reclaimed water quality requirements for agricultural irrigation according to new EU Regulation (2020/741) are reported in Table 1.
In fact, if not adequately treated, urban wastewater represents a source of microbiological health risk: its direct use in agriculture can lead to the production of contaminated food and to the spread of contamination in the environment through water leaching and runoff [5]. Moreover, the wastewater reuse can pose direct risks to workers or indirect risks to people living near application sites due to aerosolisation. It is known that numerous pathogenic microorganisms with various resistance capabilities towards environmental conditions and wastewater treatments can be transmitted through water reuse. Therefore, it is important to study pathogens that can represent a health risk in wastewater reuse.
Different studies have shown that pathogenic Campylobacter (jejuni and coli), Salmonella spp. and E. coli that produce Shiga toxins (STEC) were observed in treated wastewater [6], highlighting that irrigation with reused water could represent a source of contamination for fruit and vegetables. The Surveillance for Foodborne Diseases Outbreaks-United States annual reports from 2009 to 2015-reported Salmonella and STEC as two of the most common causes of large outbreaks frequently associated with vegetable consumption [7].
Additionally, in the EU, Campylobacter (jejuni and coli), Salmonella (mainly S. Enteritidis serovar) and STEC represent the most frequently reported causative agents of foodborne outbreaks in 2019; different cases of illness caused by Salmonella and STEC are associated with vegetables [8]. Moreover, the raw vegetables are a relevant source of risk for campylobacteriosis for consumers [5].
Wastewater reuse in agriculture may also represent a possible health risk associated with aerosol production. In fact, the irrigation method (e.g. spray irrigation) can affect the dispersion of aerosols containing pathogens such as Legionella, associated with occupational risk. Although it is known that wastewater could represent an important source of Legionella, only a few studies have focused on the occurrence of this bacterium in wastewater treatment systems [9,10]. Several outbreaks of Legionnaires' disease have been associated with aerosol production in the biological step of wastewater treatment plants, and aerosols can spread great distances from the treatment site (up to 1.6 km) [10]. Both reuse for agricultural and landscape irrigation and for cooling purposes should be considered possible sources of aerosol generation, underlining the importance of Legionella monitoring in wastewater-treated effluents [11]. The occurrence of Legionella was reported in reclaimed wastewater utilised for different purposes (e.g. agricultural use and toilet flushing) in the USA, France and Australia [10].
To improve the microbiological quality of wastewater for reuse purposes, different approaches can be used: it is possible to introduce new technologies in the secondary step of wastewater treatment (e.g. membrane bioreactors, MBRs) or optimise the performance of the final disinfection step of the wastewater treatment plant (e.g. type and amount of disinfectant). The advantages related to MBR use are a small footprint and lower chemical requirements than conventional membrane systems. However, the main disadvantage is the energy consumption largely associated with membrane fouling [12,13]. MBRs could be particularly suitable for wastewater reuse applications because this technology seems to show a greater ability to reduce some viruses (e.g. coliphage) or bacterial indicators (e.g. faecal coliforms) with respect to conventional activated sludge processes [13]. Despite these promising results, further investigation is deemed required to gain a more complete understanding, because different factors can influence the MBR performance (e.g. the nominal pore size, membrane integrity, development of a gel/cake layer on membrane surface, the adsorption of pathogens to suspended solids) [12,13]. In this context, it is of particular interest to study the abatement against different bacterial pathogens in full-scale wastewater treatment plants. Table 1 Classes of reclaimed water quality, permitted agricultural use, irrigation method and reclaimed water quality requirements for agricultural irrigation according to new EU Regulation (2020/741) [4] *If the same type of irrigated crop falls under multiple categories of the table, the requirements of the most stringent category shall apply **Drip irrigation (also called trickle irrigation) is a micro-irrigation system capable of delivering water drops or tiny streams to the plants and involves dripping water onto the soil or directly under its surface at very low rates (2-20 L/h) from a system of small-diameter plastic pipes fitted with outlets called emitters or drippers ***In the case of irrigation methods which imitate rain, special attention should be paid to the protection of the health of workers or bystanders. For this purpose, appropriate preventive measures shall be applied Minimum reclaimed water quality class A B C D Crop category (*) All food crops consumed raw where the edible part is in direct contact with reclaimed water and root crops consumed raw Food crops consumed raw where the edible part is produced above ground and is not in direct contact with reclaimed water, processed food crops and nonfood crops including crops used to feed milk-or meat-producing animals Food crops consumed raw where the edible part is produced above ground and is not in direct contact with reclaimed water, processed food crops and non-food crops including crops used to feed milk-or meat-producing animals  For the tertiary treatment of wastewater before reuse, disinfection can be carried out with chemical substances (e.g. Cl, ClO 2 , PAA) or physical treatments (e.g. UV) [14]. Chlorine dioxide (ClO 2 ) has several advantages: it is a better bactericide and virucide than sodium hypochlorite and chlorine, in relationship to its greater oxidative capacity [15]. However, physical and chemical characteristics of wastewater (temperature, pH, organic load) can influence the disinfecting capacity of ClO 2 and, consequently, the disinfectant dosage required and the contact time. Different studies comparing the removal of bacteria indicators with ClO 2 and other tertiary treatments showed that ClO 2 is more efficient than peracetic acid (PAA) and UV at high concentration (5 mg/L) [16]. At lower dosage (2 mg/L) ClO 2 and PAA presented similar abatement against bacterial and viral indicators [15].
The purpose of this study was to evaluate the role of wastewater treatment on microbiological contamination, investigating the possible risks associated with wastewater effluent reuse, with a view to new EU legislation on minimum requirements for water reuse. For this scope, pathogenic bacteria (Salmonella, STEC, thermotolerant Campylobacter, Legionella) and microbial indicators (total coliform, E. coli, enterococci and spores of Clostridium perfringens) were compared in different fullscale municipal wastewater plants equipped with MBR systems or disinfection step with ClO 2 . Moreover, a comparison between molecular and culture methods for Salmonella and Legionella was discussed.

Sampling
Samples were collected from three Italian wastewater treatment plants (WWTP1, WWTP2, WWTP3) in seven sampling periods (from January to September 2020). WWTP1 (35,000 population equivalent) and WWTP2 (36,000 population equivalent) employ wastewater pretreatment (screening, sand and oil removal), biological treatment and ultrafiltration with monolithic hollow fibre membranes (MBR) of polyvinylidene fluoride (PVDF) (nominal membrane pore size: 0.04 µm; trans membrane pressure-TMP (max): 60 kPa). WWTP3 (168,000 population equivalent) employs wastewater pre-treatment (screening, sand and oil removal), primary settling, oxidation, secondary settling, and disinfection with ClO 2 (dose applied 1.2 mg/L; contact time 1 h) before effluent discharging. The wastewater influent and the final effluent were sampled in WWTP1 and WWTP2, while in WWTP3, the effluent before and after the final disinfection treatment (final effluent) was monitored. The diagram of wastewater treatment plants and the sampling points is shown in Fig. 1.

Microbial indicators
All samples were analysed for total coliforms, enterococci, E. coli and Clostridium perfringens spores counts. Quanti-TrayTM 2000 (IDEXX Laboratories, Milan, Italy) was utilised for the quantification of coliforms, enterococci and E. coli, and the results were expressed as Log MPN/100 mL. The enumeration of C. perfringens spores was performed using a membrane filtration method according to ISO 14189:2013 [17], and the results were reported as Log CFU/100 mL. All wastewater samples were transported refrigerated and processed within 24 h of collection.

Pathogens
The presence/absence of STEC, Salmonella spp. and pathogenic Campylobacter (jejuni and coli) in wastewater samples was carried out using a PCR method with a previously reported protocol [7]. Briefly, samples (influent = 100 mL; effluent = 1 L) were concentrated by filtration and enriched in specific media. DNA extraction and purification was carried out with the DNeasy Pow-erSoil Kit (Qiagen). PCR was conducted as reported in Bonetta et al. [7,18]. Positive controls were prepared spiking an influent sample, for each sampling period, with high concentration (~ 10 6 CFU) of E. coli O157:H7 (NCTC129), S. typhimurium (ATCC14028) and Cam. jejuni (ATCC33291).
Legionella detection (spp. and pneumophila) was carried out using a quantitative molecular PCR method as previously described [19]. Briefly, 20 mL of influent and 250 mL of different effluents were concentrated by filtration; DNA was extracted from filter using the DNeasy PowerWater Kit (Qiagen) and quantified using commercial kits (iQ-Check Quanti Legionella spp. and Leg. pneumophila-BioRad). Each sample was tested in duplicate and the results were reported as the genome units (GUs) per litre of sample. The detection limit was 2500 GU/L for influent samples and 200 GU/L for effluent or predisinfected samples. The quantification limit was 20,000 GU/L for influent samples and 1600 GU/L for effluent or pre-disinfected samples.
Both Salmonella spp. and Legionella spp. were monitored with culture method in parallel with molecular detection. The culture method described in Bonetta et al. [7] was used for Salmonella spp. and results were expressed as presence/absence. The isolation and identification of Legionella was performed following ISO 11731:2017 [20]. Briefly, the effluents (concentrated by centrifugation) and the influents were analysed untreated or pretreated with heat (50 C° for 30 min) and acid (pH 2.2 for 5 min) to reduce the growth of interfering microflora, serially diluted, plated on GVPC agar and incubated for 7-10 days at 37 °C. Suspected colonies were randomly selected for subculture on α-BCYE agar and α-BCYE agar without l-cysteine. Confirmed colonies were identified using an agglutination test (Legionella latex test; Oxoid). All wastewater samples were transported refrigerated and processed within 24 h of collection. The results were reported as CFU/L; the detection limits were 2 × 10 3 and 10 2 CFU/L for influents and effluents, respectively. S. typhimurium (ATCC14028) and Leg. pneumophila (ATCC 33152) strains were used as positive control.

Physical and chemical analyses
Chemical oxygen demand (COD) was determined using the sealed tube method [21]; biochemical oxygen demand (BOD) was determined by dilution and seeding with suppression of nitrification [22]; and total suspended solids (TSS) was collected by filtration on membrane and gravimetric analysis was performed after filter drying [23]. These parameters were not available for pre-disinfected samples in the WWTP3.

Statistical analyses
The presence/absence of pathogens and concentration of bacterial indicators and Legionella (Log conversion) were statistically analysed (IBM SPSS Statistics version 26.0 for Windows).
The normal distribution was evaluated using Shapiro-Wilk test. To analyse the differences in microbial contamination between the influent and effluent samples, the effluents collected in the pre-and post-disinfection step and influent samples of WWTP1 and WWTP2 Student's t-test was applied. ANOVA and Tukey's post hoc analysis were carried out to study the effect on microbial contamination of the different treatments performed in the three WWTPs. To evaluate the relationship between concentration of indicator bacteria and the occurrence (presence/absence) of Salmonella spp., STEC and Campylobacter (jejuni and coli) in wastewater samples binary logistic regression was carried out. Spearman's correlation was used to evaluate the association between the microbiological and physicochemical parameters in each wastewater plant. Significance was evaluated within 95% confidence intervals (p ≤ 0.05).

Microbial indicators
The results of the microbial indicator concentrations in all samples investigated are reported in Fig. 2. As observed, the concentrations of E. coli, total coliforms, enterococci and spores of C. perfringens were similar in the influent samples collected from WWTP1 and WWTP2 (t-test, p > 0.05), highlighting that the microbiological characteristics of the two influents are analogous. By comparing the indicator concentrations in the effluents of the three plants, in general, the counts were higher in WWTP3 than in the other two plants. However, this difference was statistically significant only between WWTP3 and WWTP2 for all parameters, with the exception of the enterococci count (ANOVA, p < 0.05).
In both WWTP1 and WWTP2, a statistically significant reduction (t-test, P < 0.001) of all the microbiological parameters investigated was revealed in the effluent with respect to the influent, with a range of abatement of 3-5 Log and 3.5-6 Log for WWTP1 and WWTP2, respectively (Fig. 3). These results showed good abatement performance for both wastewater treatments in which MBRs were used, especially for WWTP2, although no statistically significant differences were observed between the effluents of WWTP1 and WWTP2. This behaviour could be related to the fact that the two WWTPs had MBR with the same characteristics (e.g. kind of membrane, membrane pore size, membrane material, TMP). Bolzonella and collaborators [24] also reported similar abatement values by MBR. However, other studies highlighted a higher efficiency of this treatment on domestic and industrial wastewaters, allowing reduction values up to 7 log for coliforms and E. coli and 5 log for C. perfringens [25,26]. Moreover, a further improvement of the microbial reduction was observed by adding a disinfection step with hypochlorite and/or UV, reaching a total reduction of E. coli and enterococci and a further abatement (~ 1Log) of total coliforms [27,28].
In the WWTP3, the mean reduction of microbial indicators range from 0.63 to 1.44 with values > 1 Log for E. coli and Enterococci (mean concentration of ClO 2 : 1.2 mg/L) (Fig. 3). Similar efficiency of the disinfection step with ClO 2 was observed by De Luca and collaborators in a WWTP that operated with a ClO 2 dose of 1.5 mg/L. In the same study a better performance was obtained with 2 mg/L of ClO 2 applied dose [15]. Moreover, other studies reported higher values of microbial abatement related to final disinfection with ClO 2 (range 2-5 Log for total coliforms and E. coli). However these results were obtained combining the ClO 2 disinfection step with a pre-treatment phase (e.g. filtration or adding flocculant agent) [16] or using higher doses of ClO 2 [29].
In the current Italian legislation for wastewater reuse, the regulatory reference limit for E. coli concentration is 10 CFU/100 mL (for at least 80% of samples) [30]. No samples of the effluent collected in the WWPT1 comply with this limit, underlining that, in these conditions, this treated wastewater could not be reused for irrigation, industrial or civil purposes (Additional file 1: Fig. S1). Forty-three percent of effluent samples (3/7) of WWTP2 presented values below this limit, highlighting a greater potential of the effluent of this WWTP for reuse. Since both WWTPs are equipped with MBR technology, generally suitable in the case of wastewater reuse, a thorough and careful analysis of the process conditions and the filtration efficiency of the membranes would allow us to identify useful changes to obtain a higher microbiological quality of the wastewater. In addition to membrane characteristics other factors that can affect pathogen removal by membranes should be considered and managed. In fact the development of a gel/cake layer on membrane surface, due to the accumulation of biomaterials, can modify the pathogens removal. The adsorption of pathogens to suspended solids, which depends on the characteristics of the influent water (e.g. pH, concentration of NH 3 or NaCl), can also affect membranes removal capacity. Membrane integrity is another important factor, because irregularities due to physical and chemical damage can result in a loss of pathogen removal efficiency [12,13]. Moreover, as reported in other studies [27,28], the application of an adequate final disinfection of the effluent could contribute to a further microbial reduction. However, this possibility should be carefully evaluated from an economic perspective, considering that the same MBR systems already involve higher energy and management costs than traditional activated sludge treatment.
The analysis of the results obtained in the samples from WWTP3 effluents demonstrate that only one sample complied within the limit value for E. coli, while the others showed concentrations higher than 2 log MPN/100 mL and up to 5 log MPN/100 mL, highlighting the need for a more efficient tertiary treatment (e.g. higher ClO 2 dose) or, in some cases, the requirement to secondary treatment improvement (e.g. MBR technologies).
The new EU Regulation [4] for wastewater reuse establishes different limit values for E. coli concentration according to the specific use (Table 1). Referring to this regulation, the effluent of all WWTPs, except a sample of WWTP3, complied with both the limit values for "industrial, energy and seeded crop use" (≤ 10,000 CFU/100 mL, class D) and for "drip irrigation of food crops consumed raw (with no direct contact of edible part of vegetable with reclaimed water), processed food crops and non-food crops" (≤ 1000 CFU/100 mL, Fig. 3 Mean reduction (± standard deviation) (Log CFU/100 mL or Log MPN/100 mL) of microbial indicators in the WWTPs class C). Moreover, 71% (5/7), 86% (6/7) and 14% (1/6) of all effluents analysed showed E. coli values lower than 100 CFU/100 mL (Class B), a limit established for water reused for the same food crops previously reported but "using all irrigation method". Only a few samples complied with the limit prescribed for the irrigation of all food crops consumed raw (with direct contact of edible part of vegetable with reclaimed water) and root crops consumed raw (≤ 10 CFU/100 mL, class A). Table 2 shows the results of Salmonella spp., STEC and Campylobacter (jejuni and coli) detection carried out in the three wastewater plants monitored. The presence of Salmonella spp. was reported in 100% (7/7) and 86% (6/7) of WWTP1 influents using molecular and cultural methods, respectively. In the WWTP2 Salmonella spp. was detected in 71% (5/7) of influent samples with the molecular method, while a sole sample (14%) was positive with the culture method. Additionally, Cataldo and collaborators [31] reported a similar percentage of contamination in the influent samples (from 100 to 25%). Fifty percent of pre-disinfected samples (3/6) collected in the WWTP3 were contaminated by Salmonella spp. with both methods. These results highlight a greater sensitivity of the molecular method over the culture method. This is probably related to the difficulty of isolating Salmonella on culture media due to the presence of interfering microflora, as previously reported [32] or to the use of an enrichment step for the molecular method that can aid the recovery of injured, stressed or lag-phase bacterial cells [33]. Moreover, Salmonella spp. was not detected in the effluents of WWTP1 and WWTP2 using either method, highlighting the efficacy of MBR in Salmonella abatement. In contrast, in 17% (1/6) of the WWTP3 disinfected samples, Salmonella spp. contamination was observed with both methods underlining that the disinfection step, although showing a reduction in positive samples, was not able to remove the contamination completely. No relationship between Salmonella spp. presence and microbial indicators was observed with binary logistic regression (p > 0.05); this finding suggests that the microbial parameters analysed are probably not reliable indicators of the presence of this pathogenic bacterium [6,34].

Pathogens
Italian legislation for wastewater reuse in agriculture requires the absence of Salmonella spp. in 100% of effluent analysed with the culture method. All effluents of WWTP1 and WWTP2 complied with this limit, and therefore, on the basis of these parameters and excluding high counts for E. coli, they could potentially be utilised for reuse purposes. Additionally, in the case of the WWTP3, considering the presence of Salmonella spp. in one sample in addition to the high counts of E. coli, further investigations on the efficiency of the disinfection treatment is necessary to improve microbial reduction and obtain the necessary requirements for the reuse of wastewater.
However, considering Salmonella spp. contamination, it is important to emphasise that in the new EU regulation (2020/741) on minimum requirements for water reuse [4], this parameter was deleted from the water quality requirements (Table 1); therefore, with reference to this regulation, the WWTP3 effluent could also be reused in agriculture.
As reported in Table 2, E. coli O157:H7, Shiga-like toxin (I and II) and Campylobacter (jejuni and coli) were never found in the samples analysed (influents, effluents, pre-and post-disinfection samples). Since the spread of these microorganisms in wastewater influents can reflect the local epidemiological situation and the presence of contamination sources (e.g. livestock production, food industries and slaughterhouses) [35], the absence of these bacteria could be related to a low circulation of these pathogens in the local population. The absence of STEC and pathogenic Campylobacter (jejuni and coli) in wastewater samples was also reported in other studies [6,32,36].
The presence of Legionella pneumophila sg1 was only observed with the culture method in a sole sample of the WWTP1 influent (6.3 Log CFU/L), and in the same sampling, the effluent was contaminated by Legionella spp. (3 Log CFU/L). Mosteo et al. [37] also showed low contamination by Legionella spp. in wastewater samples using culture method. Additionally, molecular analysis with qPCR reported the occurrence of Legionella spp. in all samples (influents, effluents, pre-and post-disinfection samples) collected in the three plants (Fig. 4) with a range of concentrations between 5.1 and 7.2 Log GU/L. Leg. pneumophila was detected in 86% and 57% of the WWTP1 and WWTP2 influents, respectively, and in all pre-disinfected samples of WWTP3. In contrast, among effluent samples, only some of the WWTP3 samples were contaminated by Leg. pneumophila (50% of disinfected samples). The statistical analyses confirmed the higher values of Legionella spp. and Leg. pneumophila concentrations in effluents of WWTP3 with respect to WWTP1 and WWTP2 (ANOVA, p < 0.05). It is important to highlight that the Leg. pneumophila concentrations monitored in all samples were very low or lower than the limit of quantification (LOQ), as reported in other studies [38].
For WWTP1 and WWTP2, a significant reduction in Legionella spp. and Leg. pneumophila contamination was obtained with MBR wastewater treatment (range 1.8-2 Log for Legionella spp. and 2.2-3 Log for Leg. pneumophila) (t-test, p < 0.001). In contrast, this trend was not encountered in WWTP3, showing that the reduction obtained with the disinfection step with ClO 2 was not relevant. It is important to note that in different studies conducted in WWTPs using different biological processes (including chlorination treatment), the abundance of Legionella was similar or higher in treated wastewater than in influent even at concentrations of chlorine higher (2-3 mg/L) than those used in this study.
This fact confirms that Legionella can survive within the different steps of wastewater treatment because this bacterium is able to grow within protozoa or biofilms [10,39]. Then, the significant reduction obtained in WWTP1 and WWTP2 investigated in this study highlights that the MBR treatment and the operational parameters used allow to control of Legionella spp. contamination, as was also observed for Salmonella spp. As reported above, lower Legionella contamination was observed with culture than with molecular methods in the monitored samples. This difference can be attributed to several limitations of the culture method, including the presence of abundant and/or competitive flora, the pre-treatment conditions that could inhibit Legionella growth, and the inability of this technique to detect viable but non-culturable forms (VBNCs) [10]. In fact, in wastewater samples investigated high level of interfering microflora was observed; the plates reading of untreated samples was difficult and heat or acid pre-treatment of samples was often used. The underestimation of Legionella occurrence with the culture method in wastewater samples has been reported in other studies [9,40]. On the other hand, the qPCR can instead overestimate health risks since it can also detect dead or damaged microorganisms and this deserves particular attention especially for samples that are subjected to treatments. Other aspects should be considered for the application of the molecular method in wastewater samples: despite the rapidity in the response times of the analysis, the cost of materials and equipment are high; the PCR inhibitors can be present in the samples, then appropriate internal controls should be used; moreover, there is currently non consensus on how qPCR results should be translated into the quantitative limits based on culture reported in legislation. In fact, despite its disadvantages, the culture method actually represents the gold standard to quantify viable and culturable Legionella. Additionally, in the new EU regulation for wastewater reuse, Legionella detection with the culture method was added as a water quality parameter (Legionella spp.: < 1000 CFU/L) when a risk of aerosol formation is supposed (Table 1). Referring to this limit, all effluents analysed in this study comply with EU new regulations; however, the high concentration of Legionella spp. and the presence of Leg. pneumophila in the WWTP samples analysed with qPCR underlines the need to carry out further investigation when the reuse of these effluents in agriculture is purposed.

Physical and chemical analyses
The physicochemical characteristics of the wastewater samples are shown in Fig. 5 (Additional file 1: Table S1). All values measured for BOD and almost all for TSS comply with the limits reported in the new EU regulation for reclaimed water quality class A, except for two effluents that have the TSS values required for reclaimed water quality class B (Table 1). A significant correlation was observed between microbiological parameters (total coliforms, E. coli, enterococci, C. perfringens spores, Legionella spp. and Leg. pneumophila) and physicochemical characteristics for WWTP1 and WWTP2 (Spearman correlation coefficient) (Additional file 1: Table S2), highlighting the possibility to use these parameters to support the wastewater quality evaluation and to evaluate the risk of Legionella occurrence in wastewater. It is possible to hypothesise that the physicochemical parameters monitored (TSS, BOD and COD) could be used as surrogates of microbiological indicators fate because the good correlation between parameters allows a good prediction of wastewater quality considering the reuse purpose. A positive correlation between Legionella load and COD was previously reported by Caicedo et al. [9]. The relationship between the removal of indicators and the removal of physicochemical parameters was found by Mailler et al. [41] in different wastewater samples monitored in Paris.
In contrast, no relationship was observed between the microbiological indicators and the physicochemical parameters of the WWTP3 samples, probably due to the limited number of analysed samples.
This study has various strengths: (a) it was performed in full-scale wastewater treatment plants supplied with a promising technology for the production of wastewater that can be reused in agriculture (MBR technology) or with a technology frequently used for wastewater disinfection (ClO 2 ); (b) different indicators and pathogens were investigated at the same time to evaluate the possible microbiological risk associated to the reuse purpose; (c) cultural and molecular methods were used for Salmonella spp. and Legionella to compare the results evaluating the usefulness in the monitoring programmes.
This study has also some limits: (a) the study have considered 3 WWTP S with specific characteristics (e.g. MBR pore size, ClO 2 dose), then it could be interesting to carry out further investigation considering WWTPs equipped with others MBR system or using a disinfection step improved (e.g. higher ClO 2 dose, adding of a pre-treatment) to achieve the requirement for wastewater reuse purpose; (b) the study mainly regarded bacterial contamination, but it could be interesting keeping virological risk associated to the wastewater reuse; (c) in order to study the risk associated to wastewater reuse the microbiological quality of agricultural products cultivated using treated wastewater could be also considered; (d) the use of qPCR with photoactivatable DNA intercalators (e.g. EMA or PMA) could allow to distinguish viable from dead cells improving the exposure risk evaluation to Legionella.