Comparison of UV, Peracetic Acid and Sodium Hypochlorite Treatment in the Disinfection of Urban Wastewater

One source of water contamination is the release of wastewater that has not undergone efficient treatment. The aim of this study was to evaluate the reduction obtained with sodium hypochlorite (NaClO), UV and peracetic acid disinfection treatment of Salmonella spp., pathogenic Campylobacter, STEC and bacterial indicators in three full-scale municipal wastewater plants. A general reduction in Salmonella was observed after disinfection, but these bacteria were detected in one UV-treated sample (culture method) and in 33%, 50% and 17% of samples collected after NaClO, UV and PAA disinfection treatments, respectively (PCR method). A better reduction was also observed under NaClO disinfection for the microbial indicators. Independent of the disinfection treatment, E. coli O157:H7 was not detected in the disinfected samples, whereas some samples treated with UV and PAA showed the presence of the stx1 gene. No reduction in the presence of stx2 genes was verified for any of the disinfection treatments. Campylobacter was not detected in any of the analysed samples. The overall results highlight a better reduction in microbiological parameters with a NaClO disinfection treatment in a full-scale municipal wastewater plant compared with UV and PAA. However, the results indicate that a complete and specific monitoring program is necessary to prevent a possible risk to public health.


Introduction
Water scarcity is currently one of the main challenges being faced by humans and governments; consequently, water quality protection for drinking purposes, aiming to reduce the impact on human health, represents a priority issue [1]. One of the main sources of water contamination is the release of wastewater that has not been properly treated, representing an important possible contributor to numerous pathogens and chemical contaminants [2,3]. Primary and secondary wastewater treatment typically achieves 90-99.9% enteric microflora reduction, but in some cases, this reduction may be poor. For this reason, an additional conclusive step of effluent disinfection must generally be performed [4,5]. An ideal disinfectant for this step should efficiently remove the pathogenic microorganism's maximum, avoiding the neoproduction of dangerous and undesirable by-products, and should be inexpensive and technologically compatible [6]. The chlorination is the most widely used method for the disinfection of wastewater, because it showed high effectiveness and low residual effect and cost [7]. Although this treatment inactivates different pathogens, the common application of chlorination in wastewater disinfection has led to concerns regarding the presence of disinfection by-products (such as the total organic halogen) that are cytotoxic, genotoxic and carcinogenic [8]. Thus, alternative disinfection treatments are utilised, such as ultraviolet (UV) radiation, peracetic acid (PAA) and ozonisation. The main physical procedure for the disinfection of wastewater is UV irradiation; one of the advantages of this method is that it produces nontoxic by-products [9]. However, small doses might not be sufficient to inactivate some microorganisms, and they might reverse the radiation effect through photoreactivation or dark repair. In addition, the total suspended solids and turbidity in the sewage were found to cause a lower microbiological reduction. [3,10]. PAA is a strong oxidant that represents different advantages: first, it has a large spectrum of antimicrobial activity, which is not influenced by different values of organic matter, without producing toxic and/or mutagenic by-products; moreover, it does not require dechlorination and it presents a low dependence on pH [11,12]. Although PAA can reduce the main enteric bacteria in wastewater, this disinfectant is associated with microbial regrowth and it shows lower abatement of some microorganisms (e.g., viruses and parasites) [13,14].
The information above on the main techniques applied for wastewater disinfection underlines the need to identify the most suitable processes to treat sewage. In fact, the contamination of surface waters by pathogenic microorganisms, as a consequence of inadequate wastewater disinfection, may promote waterborne disease spread, including those caused by pathogenic Campylobacter, Shiga toxin (Stx)-producing Escherichia coli (STEC) and Salmonella [15][16][17]. Although the predominant mode of Campylobacter and STEC transmission to humans is via the consumption of contaminated foods, water can also be a source of human exposure [18,19]. The largest reported waterborne outbreak of Escherichia coli O157:H7 in the United States was a consequence of a coinfection with Campylobacter jejuni, affecting 775 persons in New York State in August 1999. The epidemiological investigation has identified discharges of wastewater into river as the main source of contamination of drinking water supplies [20]. Although O157:H7 represents the most investigated strain, recent studies also highlight the role of non-O157 STEC strains as pathogens in aquatic ecosystems [21,22]. A large outbreak (1431 cases), ascribed to Campylobacter water contamination, was reported in Switzerland in 1998. The outbreak was due to pump failure producing a spill of sewage into the groundwater [23]. Recently, a campylobacteriosis outbreak (39 cases) associated with a municipal water system contaminated by wastewater was reported in Nebraska [24]. Moreover, it is also important to highlight that a possible risk for humans can also derive from the spreading of pathogenic bacteria through irrigation with surface waters contaminated by wastewater effluents. For example, according to the data provided by Centers for Disease Control and Prevention, Salmonella is the main aetiological agent involved in the foodborne disease outbreaks (~53.4%) from 2006 to 2017, and the consumption of produce was frequently associated with these outbreaks (~32.7%). The recent literature suggests that irrigation water represent a possible source of Salmonella contamination in produce, highlighting its possible role as a transmission vehicle [25].
The aim of this study was to evaluate the reduction obtained with sodium hypochlorite, UV and peracetic acid disinfection treatment of the most important zoonotic bacterial pathogens (Salmonella spp., pathogenic Campylobacter, and STEC) and typical bacterial indicators of faecal contamination in three different full-scale municipal wastewater plants.

Bacterial Indicators
The presence of E. coli, coliform bacteria, enterococci, and C. perfringens spores was investigated, comparing the reduction obtained with the three different disinfection treatments (WWTP1 with NaClO, WWTP2 with UV, WWTP3 with PAA). Moreover, the relationship between bacterial concentrations and pathogens was evaluated. The mean concentrations of the faecal indicators found in the three WWTPs are reported in Table S1 (Supplementary Material).
Generally, in the non-disinfected effluents (E), the concentrations of the indicators were higher in WWTP1 and WWTP2 than in WWTP3 (p > 0.05). Moreover, in all WWTPs, the bacterial levels observed in non-disinfected effluent were similar to those reported in other studies [3,26,27]. Furthermore, the decreasing trend of the mean concentrations of the four indicators in the three WWTPs was total coliforms > E. coli > Enterococci > C. perfringens. The mean concentrations of coliform bacteria and C. perfringens spores were similar in the non-disinfected effluent of the three WWTPs (ANOVA p > 0.05) ( Table S2, Supplementary Material). A difference between the mean concentration of E. coli and enterococci in the non-disinfected effluents of WWTP1 and WWTP2 with respect to WWTP3 was observed, but this difference was statistically significant only between WWTP1 and WWTP3 (ANOVA p < 0.005 for E. coli and p < 0.05 for enterococci) (Table S3, Supplementary Material).
Moreover, ANOVA highlighted that the bacterial load in the WWTPs was similar between the seasons (p > 0.05).
The decreasing trend of the mean concentration of the four indicators in the disinfected effluents (DE) of the three WWTPs was C. perfringens ≥ Enterococci > total coliforms > E. coli (Table S3, Supplementary Material). This trend was attributable to the different sensitivities of the four indicators to the disinfection treatments applied, as reported by other authors [27][28][29]. The concentrations of the four indicators were lower in WWTP1 than in WWTP2 and WWTP3. The ANOVA and post-hoc tests highlight a significant difference between the mean concentration of total coliforms, E. coli and enterococci in the disinfected effluents of WWTP1 vs. WWTP2 and WWTP3. Otherwise, the mean concentrations of C. perfringens spores were similar in the disinfected effluent of the three WWTPs (ANOVA p > 0.05).
Moreover, a significant difference in the concentrations of the four indicators between the E and DE of each WWTP was observed, with the exception of C. perfringens spores in the WWTP3. Therefore, all three disinfection treatments were useful in reducing the microbial concentration in the final effluent (Table S2).
Comparing the total mean reduction in all the indicators for each WWTP (Figure 1), a significant difference was observed among the three WWTPs (ANOVA p < 0.0005), particularly between WWTP1 and WWTP2 (post hoc p < 0.0005) and between WWTP1 and WWTP3 (post hoc p < 0.0005).
tionship between bacterial concentrations and pathogens was evaluated. The mean concentrations of the faecal indicators found in the three WWTPs are reported in Table S1 (Supplementary Material).
Generally, in the non-disinfected effluents (E), the concentrations of the indicators were higher in WWTP1 and WWTP2 than in WWTP3 (p > 0.05). Moreover, in all WWTPs, the bacterial levels observed in non-disinfected effluent were similar to those reported in other studies [3,26,27]. Furthermore, the decreasing trend of the mean concentrations of the four indicators in the three WWTPs was total coliforms > E. coli > Enterococci > C. perfringens. The mean concentrations of coliform bacteria and C. perfringens spores were similar in the non-disinfected effluent of the three WWTPs (ANOVA p > 0.05) (Table S2, Supplementary Material). A difference between the mean concentration of E. coli and enterococci in the non-disinfected effluents of WWTP1 and WWTP2 with respect to WWTP3 was observed, but this difference was statistically significant only between WWTP1 and WWTP3 (ANOVA p < 0.005 for E. coli and p < 0.05 for enterococci) (Table S3, Supplementary Material).
Moreover, ANOVA highlighted that the bacterial load in the WWTPs was similar between the seasons (p > 0.05).
The decreasing trend of the mean concentration of the four indicators in the disinfected effluents (DE) of the three WWTPs was C. perfringens ≥ Enterococci > total coliforms > E. coli (Table S3, Supplementary Material). This trend was attributable to the different sensitivities of the four indicators to the disinfection treatments applied, as reported by other authors [27][28][29]. The concentrations of the four indicators were lower in WWTP1 than in WWTP2 and WWTP3. The ANOVA and post-hoc tests highlight a significant difference between the mean concentration of total coliforms, E. coli and enterococci in the disinfected effluents of WWTP1 vs. WWTP2 and WWTP3. Otherwise, the mean concentrations of C. perfringens spores were similar in the disinfected effluent of the three WWTPs (ANOVA p > 0.05).
Moreover, a significant difference in the concentrations of the four indicators between the E and DE of each WWTP was observed, with the exception of C. perfringens spores in the WWTP3. Therefore, all three disinfection treatments were useful in reducing the microbial concentration in the final effluent (Table S2).
Comparing the total mean reduction in all the indicators for each WWTP (Figure 1), a significant difference was observed among the three WWTPs (ANOVA p < 0.0005), particularly between WWTP1 and WWTP2 (post hoc p < 0.0005) and between WWTP1 and WWTP3 (post hoc p < 0.0005).  The mean removal of each microbial indicator by the three WWTPs is shown in Figure 2: a lower reduction was found in the three WWTPs for C. perfringens spores, without any significant differences among the three disinfection treatments; otherwise, E. coli, coliform bacteria and enterococci reached a high abatement in all plants with significant differences related to the application of the different treatments (NaClO, UV and PAA) (ANOVA p < 0.0005). The mean removal of each microbial indicator by the three WWTPs is shown in Figure 2: a lower reduction was found in the three WWTPs for C. perfringens spores, without any significant differences among the three disinfection treatments; otherwise, E. coli, coliform bacteria and enterococci reached a high abatement in all plants with significant differences related to the application of the different treatments (NaClO, UV and PAA) (ANOVA p < 0.0005). In particular, the reduction in coliforms, E. coli and enterococci in WWTP1 because of NaClO disinfection was significantly greater than those reached in WWTP2 (post hoc p < 0.0005, p < 0.005 and p < 0.005, respectively) and WWTP3 (post hoc p < 0.0005, p < 0.0005 and p < 0.0005, respectively), in which disinfection was made with UV and PAA, respectively. Therefore, NaClO disinfection is confirmed as the most effective treatment in terms of indicator microorganism reduction. It is important to emphasize that all three disinfection processes allow values below the regulatory reference limit for E. coli concentration (5 × 10 3 CFU/100 mL) to be reached with reference to wastewater discharge into surface waters in Italy [30]. However, the non-disinfected effluents of all plants always presented concentrations above this limit, underlining the need for a disinfection step to reduce the microbial impact of wastewater effluents on the receiving surface water. Considering that Italian legislation for the reuse of wastewater effluents establishes a legal limit of 1 Log CFU/100 mL for E. coli concentration [31], disinfection with NaClO always reached this goal, except in July, which was probably due to the particularly high concentration of E. coli in the non-disinfected effluent. In contrast, the effluents disinfected with UV and PAA did not comply with the limit, except in two samples disinfected by UV (January and July) and in one sample disinfected by PAA (January).
Pearson's correlation analysis shows a positive correlation, even if not significant, between the dosage of NaClO applied in the WWTP1 and the reduction in coliforms (r = 0.771), of E. coli (r = 0.756) and the total mean abatement (r = 0.626). Additionally, for NaClO contact times, positive but not significant correlations were found, with E. coli (r = 0.761), C. perfringens (r = 0.608) and the total mean abatement (r = 0.631), while a positive and significant correlation with coliform bacteria was found (r = 0.898; p < 0.05). These results confirm the sensitivity of coliforms and E. coli to NaClO and show that the reduction in C. perfringens was mainly dependent on the contact time rather than on the dosage used. In particular, the reduction in coliforms, E. coli and enterococci in WWTP1 because of NaClO disinfection was significantly greater than those reached in WWTP2 (post hoc p < 0.0005, p < 0.005 and p < 0.005, respectively) and WWTP3 (post hoc p < 0.0005, p < 0.0005 and p < 0.0005, respectively), in which disinfection was made with UV and PAA, respectively. Therefore, NaClO disinfection is confirmed as the most effective treatment in terms of indicator microorganism reduction. It is important to emphasize that all three disinfection processes allow values below the regulatory reference limit for E. coli concentration (5 × 10 3 CFU/100 mL) to be reached with reference to wastewater discharge into surface waters in Italy [30]. However, the non-disinfected effluents of all plants always presented concentrations above this limit, underlining the need for a disinfection step to reduce the microbial impact of wastewater effluents on the receiving surface water. Considering that Italian legislation for the reuse of wastewater effluents establishes a legal limit of 1 Log CFU/100 mL for E. coli concentration [31], disinfection with NaClO always reached this goal, except in July, which was probably due to the particularly high concentration of E. coli in the non-disinfected effluent. In contrast, the effluents disinfected with UV and PAA did not comply with the limit, except in two samples disinfected by UV (January and July) and in one sample disinfected by PAA (January).
Pearson's correlation analysis shows a positive correlation, even if not significant, between the dosage of NaClO applied in the WWTP1 and the reduction in coliforms (r = 0.771), of E. coli (r = 0.756) and the total mean abatement (r = 0.626). Additionally, for NaClO contact times, positive but not significant correlations were found, with E. coli (r = 0.761), C. perfringens (r = 0.608) and the total mean abatement (r = 0.631), while a positive and significant correlation with coliform bacteria was found (r = 0.898; p < 0.05). These results confirm the sensitivity of coliforms and E. coli to NaClO and show that the reduction in C. perfringens was mainly dependent on the contact time rather than on the dosage used.
No correlation (Pearson) was found with the dosage and contact time of PAA applied in the WWTP3. No relationship between the parameters of UV treatment and the microbial indicators can be determined because the power of UV lamps and the contact time are fixed.

Pathogenic Bacteria
The results of the detection of Salmonella spp., E. coli O157:H7 and Campylobacter carried out in the three full-scale municipal wastewater plants are reported in Table 1.  Salmonella spp. were observed in all the analysed effluents using molecular methods, except for the samples collected in WWTP3 in January 2018, while they were seen in 66.6% of the non-disinfected effluents (E) with the culture method. Analogous percentages of contamination were also reported in other studies [28,32]. The presence of Salmonella in the effluents before disinfection treatment underlines that conventional municipal wastewater treatments are not able to remove these bacteria. This finding underlines the importance of using tertiary disinfection treatment to avoid the spread of enteric pathogens in the environment and to prevent a possible risk to public health [4]. Considering the different disinfection treatments, a generally consistent reduction in Salmonella contamination was observed; in particular, with the culture-based method, Salmonella was detected in a sole sample of the WWTP2 (May 2017), whereas with the PCR method, Salmonella was detected in 33% of samples (2/6) collected after NaClO disinfection, in 50% (3/6) after UV treatment and in 17% (1/6) after PAA disinfection. These results highlight that none of the disinfection treatments was able to remove Salmonella contamination completely in the final effluent; however, a better reduction was observed with PAA and NaClO disinfection with respect to UV. The same trend was also observed for the microbial indicators. The efficacy of NaClO and PAA in reducing Salmonella contamination has been reported in other studies [33,34], where the performances of different WWTP disinfection technologies were evaluated. In line with our results, Veschetti and collaborators [32] reported a similar bactericidal power of PAA and NaClO against Salmonella and other microorganisms. The lower disinfection of UV treatment with respect to PAA and NaClO could be due to the photoreactivation mechanism of Salmonella and other pathogenic bacteria. In fact, the potential regrowth and repair of pathogenic bacteria (photoreactivation) in UV-disinfected wastewater was reported as a drawback of the real application of this process [10].
Comparing the results obtained with the molecular and culture methods for Salmonella detection, a lower contamination was observed in the effluent before disinfection as well as after the treatments using the culture method. These results could be related to the difficulty to isolate the strain from XLD agar because of the presence of interfering microflora (e.g., Proteus mirabilis). This methodological problem should be taken into account when the microbiological risk associated with the reuse of wastewater effluent is estimated, considering that Italian regulation prescribes Salmonella spp. absence evaluated by the culture method [31].
The results of the PCR analyses showed that 50% (3/6) of the WWTP1 and WWTP3 effluents were positive for E. coli O157:H7, while only one of the WWTP2 non-disinfected samples (17%) was contaminated with this pathogenic microorganism. Independent of the disinfection treatment applied, E. coli O157:H7 was not detected in any of the examined disinfected samples, demonstrating the effectiveness of removal of this microorganism by the tertiary disinfection treatments investigated.
A total of four (4/6 or 67%), four (4/6 or 67%) and five (5/6 or 83%) effluents before disinfection showed the presence of amplicons corresponding to the stx1 gene (WWTP1, WWTP2 and WWTP3, respectively). The stx2 gene was detected only in one effluent (E) in WWTP2 (January 2018). The presence of amplicons corresponding to the stx1 and stx2 genes in effluents before disinfection was always associated with the O157:H7 serotype or other STECs, as reported in Table 1. After disinfection, in WWTP1, no samples revealed the presence of the stx1 amplicon, while WWTP2 and WWTP3 showed two (33%, 2/6) and one (17%, 1/6) sample with the stx1 gene, respectively. No reduction in the presence of stx2 genes was verified for the disinfection treatments; in fact, two (2/6 or 33%), one (1/6 or 17%) and one (1/6 or 17%) effluent after disinfection revealed the presence of amplicons corresponding to the stx2 gene in WWTP1, WWTP2 and WWTP3, respectively. As reported in Table 2, three samples (DE WWTP1 and WWTP3 September 2017 and DE WWTP1 May 2018) tested positive for stx genes, but they were not associated with the presence of the STEC serogroups investigated in this study, in agreement with the results obtained in our previous studies [28,35]. This finding was probably related to the presence of bacteriophages carrying stx1/stx2 genes, since their role in the dissemination of such sequences among STEC and Shigella strains and among other waterborne bacteria is well known [36,37]. Another reason for the abovementioned stx1/stx2 amplification products could be the presence of STEC serogroups other than those targeted in this research.
It is also important to consider that the presence of stx genes is essential to infection; however, other virulence factors (e.g., the eae gene) could be involved [38]. In our study, the eae gene (coding for the virulence factor intimin) was revealed in only one sample of disinfected effluent, but in this sample the stx1/stx2 genes were absent.
Campylobacter were not detected in any of the samples analysed because no amplicons corresponding to genus-specific 16S rRNA (Campylobacter spp.) and species-specific mapA and ceuE genes (for C. jejuni and C. coli species) were detected. Considering the absence of contamination by pathogenic Campylobacter, no conclusions about the reduction in the different disinfection treatments can be drawn.
To verify the correlation between the presence of emerging pathogens and the counts of faecal indicators such as E. coli, coliforms, enterococci and C. perfringens spores, a logistic binary regression analysis was performed. No associations between the presence of O157:H7 genes, the eae gene (intimin), stx1 gene (Shiga-like toxin I), stx2 gene (Shiga-like toxin II), invA gene (Salmonella spp.) and the counts of microbiological indicators (E. coli, enterococci, C. perfringens spores and coliforms) were observed (p > 0.05). Moreover, no relationships were observed between Salmonella spp. contamination revealed by the culture method and faecal indicators or other pathogens (p > 0.05). These results suggest that the common bacterial indicators of faecal contamination in municipal wastewater samples seem to not be reliable indicators of pathogenic bacteria presence. These considerations are in agreement with other studies on relationship between bacterial indicators and pathogenic bacteria, including Salmonella spp. and STEC [39,40].  To conclude, the overall results obtained underline a better reduction in microbiological parameters monitored using sodium hypochlorite as a disinfection treatment in a full-scale municipal wastewater plant with respect to UV and PAA. However, pathogen detection with molecular methods revealed the presence of Salmonella contamination and stx2 genes in effluents disinfected with NaClO. This finding highlights the need for a complete and specific monitoring program to prevent possible risks to public health, also considering the lack of correlation between pathogens and microbial indicators.

Sampling
Effluents before and after disinfection treatment were collected during six sampling periods (September 2017, November 2017, January 2018, March 2018, May 2018 and July 2018) from three Italian wastewater treatment plants (WWTP1-population equivalent of 34000, WWTP2-population equivalent of 8000, and WWTP3-population equivalent of 8000). The different wastewater treatment plants are characterized by a disinfection step with NaClO (15% w/w), UV lamp and PAA (15% w/w) for WWTP1, WWTP2 and WWTP3, respectively. Chemical-physical characteristics of non-disinfected wastewater are reported in Table 3, and information about disinfection conditions in the sampling period is reported in Table 4. After the sampling, wastewater samples were maintained at +4 • C and analyzed within 24 h.

Microbiological Analyses for Pathogen Detection
Effluents before and after disinfection were tested for pathogen detection. During each sampling, a raw sewage sample spiked with a high concentration of pathogens (~10 6 CFU/100 mL) was used as a positive control. The concentration, enrichment, DNA extraction and PCR/real-time PCR steps were carried out as reported in our previous studies [35,41,42]. Salmonella spp. detection was also evaluated using the culture method. Briefly, samples were pre-enriched in peptone water (Oxoid), enriched in selective media (Rappaport Vassiliadis Broth, RV, Oxoid) and streaked on specific media (Xylose Lysine Deoxycholate agar Oxoid) [43]. The identification of selected bacterial colonies was carried out using an API ® 20E kit (BioMerieux, Marcy L'Etoile, France).

Microbiological Analyses for the Detection of Microbial Indicators
E. coli, enterococci, Clostridium perfringens spores and coliforms were analysed in all samples. In brief, the membrane filtration method was used to process wastewater samples for C. perfringens enumeration, as reported by ISO 14189:2013 [44]. Wastewater samples were assayed for E. coli, coliforms, and enterococci with a commercial semiautomated quantification method (Quanti-TrayTM 2000, IDEXX Laboratories, Milan, Italy). With Quanti-Tray, a 100 mL sample was added to a specific substrate for each microorganism, and the mixture was divided into 51 wells and incubated at a specific temperature. Then the standard method, based on the Most Probable Number (MPN) approach, was used to determine the number of bacteria in the original sample. The Colilert-18 test uses a growth substrate with specific indicators (ONPG and MUG) to detect coliforms, and E. coli. Coliforms use their β-galactosidase enzyme to metabolize ONPG and change it from colorless to yellow. E. coli use β-glucuronidase to metabolize MUG and create fluorescence. The Enterolert Test uses another growth substrate to detect enterococci. When enterococci utilize their ß-glucosidase enzyme to metabolize Enterolert's nutrient-indicator (4-methylumbelliferyl ß-D-glucoside) the sample fluoresces, the number of positive wells are counted and referred to the MPN table provided to obtain the Most Probable Number of bacteria (MPN/100 mL of sample). This method has been approved by U.S. EPA and it has been included in Standard Methods for Examination of Water and Wastewater.

Statistical Analyses
IBM SPSS Statistics version 26.0 (IBM, Segrate, Italy) for Windows was used for statistical analyses. The relationship between the indicator bacteria load (log 10) and the presence/absence of pathogens was carried out with binary logistic regression. The reduction in indicator load using the different disinfection treatments was analyzed with ANOVA and Tukey's post-hoc tests. The association between the disinfection conditions (quantity and contact time) and the reduction in indicators was performed using Pearson's correlation.