key: cord-0747751-138lcl6v
authors: Chen, Yan-huan; Yan, Cheng; Yang, Ya-fei; Ma, Jia-xin
title: Quantitative microbial risk assessment and sensitivity analysis for workers exposed to pathogenic bacterial bioaerosols under various aeration modes in two wastewater treatment plants
date: 2020-10-01
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2020.142615
sha: c86b982a31c47ae4f8015e8fa4de7e915fe01488
doc_id: 747751
cord_uid: 138lcl6v

Wastewater treatment plants (WWTPs) could emit a large amount of bioaerosols containing pathogenic bacteria. Assessing the health risks of exposure to these bioaerosols by using quantitative microbial risk assessment (QMRA) is important to protect workers in WWTPs. However, the relative impacts of the stochastic input variables on the health risks determined in QMRA remain vague. Hence, this study performed a Monte Carlo simulation-based QMRA case study for workers exposing to S. aureus or E. coli bioaerosols and a sensitivity analysis in two WWTPs with various aeration modes. Results showed that when workers equipped without personal protective equipment (PPE) were exposed to S. aureus or E. coli bioaerosol in the two WWTPs, the annual probability of infection considerably exceeded the U.S. EPA benchmark (≤10E-4 pppy), and the disease burden did not satisfy the WHO benchmark (≤10E-6 DALYs pppy) (except exposure to E. coli bioaerosol for disease health risk burden). Nevertheless, the use of PPE effectively reduced the annual infection health risk to an acceptable level and converted the disease health risk burden to a highly acceptable level. Referring to the sensitivity analysis, the contribution of mechanical aeration modes to the variability of the health risks was absolutely dominated in the WWTPs. On the aeration mode that showed high exposure concentration, the three input exposure parameters (exposure time, aerosol ingestion rate, and breathing rate) had a great impact on health risks. The health risks were also prone to being highly influenced by the various choices of the dose–response model and related parameters. Current research systematically delivered new data and a novel perspective on the sensitivity analysis of QMRA. Then, management decisions could be executed by authorities on the basis of the results of this sensitivity analysis to reduce related occupational health risks of workers in WWTPs.

Wastewater treatment plants (WWTPs) could emit a large amount of bioaerosols containing pathogenic bacteria (Szyłak-Szydłowski et al., 2016) . Compared with other workers, workers in WWTPs have a particularly higher prevalence of the so-called -sewage worker's syndrome,‖ characterized by fatigue, headache, dizziness, gastrointestinal symptoms, and respiratory symptoms (Hung et al., 2010) . These symptoms could be caused by work-related exposure to various bacterial bioaerosols that were liberated in wastewater treatment processes (Kowalski et al., 2017) .

Staphylococcus aureus and Escherichia coli bioaerosols, which had been frequently found in domestic wastewaters, are widely used as target indicator pathogens (Ikehata, 2013; Szyłak-Szydłowski et al., 2016; Shi et al., 2018; Kozajda et al., 2019) . These bacteria from wastewater or sludge can infect people through inhalation (Szyłak-Szydłowski et al., 2016) . Direct exposure to these bioaerosols causes gastrointestinal infection through bioaerosol capture in the upper respiratory tract by inhalation, where pathogenic bacterial bioaerosols move by ciliary action and pass into the digestive tract through the pharynx (Peccia et al., 2008) . In general, the exposure of humans to WWTPs with pathogenic bacterial bioaerosols has significant health risks (Kozajda et al., 2019) . Therefore, assessing the health risks of exposure to pathogenic bacterial bioaerosols is important to protect workers in WWTPs. In addition, the aeration mode in WWTPs and using personal protective equipment (PPE) Health risk is usually quantified by the annual probability of infection (P (a)inf ) and disease burden (DB) (Haas et al., 2014) . The P (a)inf and DB of bioaerosol exposure could be estimated by quantitative microbial risk assessment (QMRA) (Haas et al., 2014; Shi et al., 2018; Esfahanian et al., 2019) . QMRA commonly follows four classical working steps: (a) hazard identification, (b) exposure assessment, (c) doseresponse assessment, and (d) risk characterization (Haas et al., 2014) . Moreover, QMRA is often estimated from Monte Carlo simulations to assess the range and likelihood of the health risk quantitatively (Lim et al., 2013; Shi et al., 2018; Liu et al., 2019) . Furthermore, for risk characterization, the two most widely used health risk benchmarks are the acceptable annual infection risk level proposed by the U.S. EPA (≤10E−4 pppy) (2005) and the acceptable disability-adjusted life years (DALYs) by the WHO (≤10E−6 DALYs pppy) (2008) . They are widely used in interpreting the magnitude of risk assessment outcomes (Lim et al., 2015; Fuhrimann et al., 2016; Shi et al., 2018) . These two benchmarks were built around the concept of health-based targets that were grounded on well-defined health metrics (e.g., DALYs) and a level of tolerable health burdens (Fuhrimann et al., 2016) .

improve the understanding and interpretation of the QMRA framework in order to extent of its analysis methodology; and (c) recognize input variable gaps and then prioritize future research priorities (Tesson et al., 2020) . Haas et al. (2017) demonstrated health risk of Ebolavirus for sewer worker with or without PPE from inhalation exposure and sensitivity analysis from Monte Carlo simulation. Kowalski et al. (2017) analyzed the emission characterization of the bacteria and fungi bioaerosols collected in different aeration modes of WWTPs in Poland. Carducci et al. (2018) reported that the health risk for workers in WWTPs exposed to the human adenovirus (HAdV) was estimated by QMRA and the sensitivity analysis was employed to examine the impact of input parameters (breathing rate and concentration) on health risk. However, given the ranking, significance, and contribution of these relative impacts remain vague, a series of open questions have been raised about the QMRA and its sensitivity analysis for stochastic input variables associated with workers using PPE and exposing under various aeration modes in WWTPs. Therefore, this research systematically investigates a Monte Carlo simulation-based QMRA case study for workers exposing to S. aureus and E. coli bioaerosols in two WWTPs. After that, the health risks (P (a)inf and DB) of the workers without or with PPE exposed to bioaerosols under various aeration modes in two WWTPs were discussed. Then, it focuses on the rank correlation coefficient values and contribution to variance of each input variable in QMRA which were assessed by sensitivity analysis. The current J o u r n a l P r e -p r o o f Journal Pre-proof research can enrich the knowledge bases of the sensitivity analysis of QMRA for workers with PPE exposed to bioaerosols under various aeration modes in WWTPs and then provide an advanced understanding of the rank correlation coefficient values and contributions to variance of each input variable in QMRA framework. These results can inform efforts to establish rational management recommendations for reducing occupational health risks of workers in WWTPs.

This study was performed in two WWTPs (plants A and B), which were located in the central part of P.R. China. Their drainage pipe systems were similar. The collected domestic wastewater (occasionally mixed with a little industrial wastewater) was distributed into the WWTP by a series of variable-frequency pump stations. Plant A had a parallel wastewater treatment system equipped with a rotating disc aeration tank (RD) and a microporous aeration tank (M), treating 50,000 tons of wastewater per day, respectively. Similarly, plant B was also a parallel system. It had an inverted umbrella aeration tank (IU) and a microporous aeration tank (M), treating 100,000 tons of wastewater per day, respectively. Thus, there were three modes for aeration tanks (RD, IU, and M) in this research.

(FA-1, Hongchangxin Inc., Beijing, China) (Hung et al., 2010) . Sterile agar media Egg-Yolk Mannitol Salt Agar Base and MacConkey-Agar-Medium were used as the collection media for culturing and colony enumeration of S. aureus and E. coli, respectively (Oppliger et al., 2005; Szyłak-Szydłowski et al., 2016; Nasir et al., 2018; Wang et al., 2019) . A 27 mL aliquot of this sterile agar media (autoclaved at 121 ℃ for 15 min) was pipetted into sterile glass Petri dishes equipped with the cascade impactor (Jahne et al., 2015; Jahne et al., 2016) .

The sampling point was set at 1.5 m above each aeration tank's ground (Szyłak-Szydłowski et al., 2016) . The cascade impactor was operated for 10 min at a flow rate of 28.3 L/min (Hung et al., 2010; Kowalski et al., 2017) . Each stage of the Andersen six-stage cascade impactor was decontaminated with 75% alcohol before and after use for air sampling on site (Hung et al., 2010) . All samples were in triplicate and transported to the laboratory in a cold box before being cultivated in incubators for 24-48 h at 37 ℃.

After cultivation, the samples were enumerated as colony-forming unit (CFU) by using an automatic colony enumeration instrument (HICC-B, Wanshen Inc., Hangzhou, China). The positive hole method was used to correct and then obtain the actual number of colonies measured at the each Petri dish stage on the basis of the enumeration results (Hung et al., 2010; Delort et al., 2017) . Bioaerosol concentrations of S. aureus and E. coli in CFU/m 3 were estimated by dividing the number of colonies in CFU by the sampled air volume in m 3 (Hung et al., 2010) . Then, the bioaerosol concentration was the sum of the concentrations of the six Petri dish stages of the Andersen six-stage cascade impactor (Katsivela et al., 2017) . The indicator pathogens of concern in this research were S. aureus and E. coli bioaerosols in the two WWTPs. So, the workers in the WWTP aeration tanks were exposed to serious S. aureus and E. coli bioaerosols-related health risks.

[ Table 1 inserts here]

[ Figure 1 inserts here]

The parameters and flow chart for the exposure assessment referring to the QMRA calculation framework are presented in Table 1 and Figure 1 , respectively. This research had eight exposure scenarios ( Fig. 1) : (a) workers without PPE exposed to S. aureus bioaerosol in plant A, (b) workers without PPE exposed to S. aureus bioaerosol in plant B, (c) workers with PPE exposed to S. aureus bioaerosol in plant A, (d) workers with PPE exposed to S. aureus bioaerosol in plant B, (e) workers without PPE exposed to E. coli bioaerosol in plant A, (f) workers without PPE exposed to E. coli bioaerosol in plant B, (g) workers with PPE exposed to E. coli bioaerosol in plant A, and (h) workers with PPE exposed to E. coli bioaerosol in plant B. The exposure concentrations (ec) of S. aureus and E. coli bioaerosols are calculated and shown in respirator at all times (i.e., workers with PPE) (Haas et al., 2017) .

The dose of pathogens (Dose) per person per day was calculated in Equation (1) (Dungan, 2014; Jahne et al., 2015; Haas et al., 2017) :

where 

where P (d)inf is the estimated daily probability of infection, and k is the parameter of the model (Table 1) . 

Risk characterization was carried out on the basis of the contaminant concentration to which individuals were exposed. Annual probability was estimated considering the number of exposure events per year with Equation (4) (Haas et al., 2014; Salesortells et al., 2014) :

where P (a)inf is the annual probability of infection per person per year (pppy), and n is the annual exposure frequency (Table 1) .

For S. aureus bioaerosol, the probability of infection was assumed equal to the probability of illness (P ill/inf = 1). The probability of illness, as a conditional of infection, was calculated in Equation (5) (Busgang et al., 2018; Carducci et al., 2018) :

where P (a)ill is the annual probability of illness, and P ill/inf is the specific conditional probability of illness given an infection (i.e., prevalence) (Table 1) .

For E. coli bioaerosol, the exponential dose-response model was used as a doseillness model to calculate the probability of illness, which was defined in Equation (6) ( Shi et al., 2018) :

where P (a)ill is the annual probability of illness, and k is the parameter of the model, which are listed in Table 1 .

The specific potential disease burden attributable to illness caused by exposure to S.

aureus or E. coli bioaerosol was estimated in Equation (7) 

where DB is the disease burden and expressed in DALYs pppy, and HB is the health burden and expressed in DALYs per illness case (DALYs/case) (Table1).

Monte Carlo simulation was used to represent the propagation of variability in QMRA (Lim et al., 2015) . It was run with 10,000 trials by using Oracle Crystal Ball [ Figure 2 inserts here]

3.1 Dose-response assessment and risk characterization Figure 2 demonstrates the annual infection risks (P (a)inf ) and the disease burdens (DB) that were estimated from the Monte Carlo simulations with 10,000 iterations under the eight exposure scenarios where workers (without or with PPE) were exposed to S. aureus or E. coli bioaerosols in the two WWTPs.

For exposing to S. aureus bioaerosol, the P (a)inf of the workers in plant A were always much higher than that of the workers in plant B (Fig. 2a) . This finding could be explained by the theory that the different aeration modes between the two WWTPs lead to huge differences in the concentration of S. aureus bioaerosol emissions, which would largely affect the annual infection health risks for workers (Haas et al., 2014; Dungan, 2014; Jahne et al., 2015) . Nevertheless, the P (a)inf of the workers without PPE in plants A and B both considerably exceeded the U.S. EPA benchmark (≤10E-4 pppy) (Fig. 2a) . However, the P (a)inf of the workers with PPE in plant A (median = 6.04E-04) was on the same order of magnitude as the benchmark, and the P (a)inf of the workers with PPE in plant B clearly satisfied the benchmark. These results indicated that using J o u r n a l P r e -p r o o f Journal Pre-proof PPE can effectively reduce the annual infection health risks of S. aureus bioaerosol to an acceptable level (Ikehata, 2013; Hass et al., 2017; Carducci et al., 2018) .

For E. coli bioaerosol, the P (a)inf of the workers without or with PPE in plant A slightly differed from that of the workers in plant B (Fig. 2a) . Furthermore, the P For S. aureus bioaerosol, the DB of the workers in plant A was much higher than that of the workers in plant B (Fig. 2b) Therefore, wearing of PPE improved the disease health risk burden of workers exposed to S. aureus bioaerosol from low acceptable level to high acceptable level. (Shi et al., 2018) . Thus, even without PPE, the disease health risk burden of the workers exposed to E. coli bioaerosol was still acceptable. Similar result had been reported. Shi et al. (2018) found that even in the worst-case scenario, where all E.

coli bioaerosols were assumed to be pathogenic, the health risks were still far below the benchmark.

What was noteworthy was that, as expected, the health risks (P (a)inf and DB) of the workers exposed to S. aureus bioaerosol with PPE reduced by approximately two orders of magnitude compared with those of the workers without PPE in plants A and B. This result was because the N-95 respirators utilized in this research were engineered to filter at least 95% of the particles that would be inhaled (Hass et al., 2017) . The results of the reduction of health risk of workers with PPE exposed to E.

coli bioaerosol were similar.

thought to be less readily available. Lim et al. (2015) put forth that the U.S. EPA P (a)inf benchmark is regionally bounded because it was proposed according to the disease surveillance data only in the U.S. Therefore, this benchmark might not be representative of the whole world. Moreover, the WHO DB benchmark should be treated cautiously in a similar manner to the U.S. EPA P (a)inf benchmark, and these two indicators ought to be used as complements rather than opposites in health risk assessment (Lim et al., 2015) . In addition, the U.S. EPA P (a)inf benchmark and the WHO DB benchmark are considered to be overly conservative (Lim et al., 2015) .

In this research, the P (a)inf and DB were calculated by using different dose-response models for the QMRA of S. aureus and E. coli bioaerosols (Shi et al., 2018) . For S.

aureus bioaerosol QMRA, the metrics used for P (a)inf and DB were directly related to each other, and the DB was calculated via P (a)inf and DALYs (Havelaar et al., 2012; Busgang et al., 2018) . By contrast, uncorrelation of the dose-response models for E.

coli bioaerosol QMRA led to the variability of the health risk calculations. P (a)inf was calculated using the Beta-Poisson dose-response model (Equations (3) and (4)), and the DB was calculated using the exponential dose-response model (Equations (6) and (7)). Thus, this research implied that the health risks (P (a)inf and DB) were prone to being highly influenced by the various dose-response models of choice. In general, accurate health risk estimation called for additional field studies and clinical infection data (Shi et al., 2018) . But there remain also need to understand that an efficient and rigorous validation of the dose-response model and its relevant parameters for QMRA is warranted (Haas, 2015) .

[ Figure 3 inserts here] (Figs. 4b and 4d ). Moreover, the contribution to variance of the exposure concentration for the workers on the M aeration tank exerted minor effect on the health risk with fraction >10% (Figs. 4b and 4d ). The three input exposure parameters (exposure time, aerosol ingestion rate, and breathing rate) for the workers on the IU aeration tank all showed a slightly higher ranking than those on the M aeration tank When the PPE employed workers exposing to the mechanical aeration tanks in plants A and B, the input variable -removal fraction by employing with PPE‖ contributed the second ranking for health risks. This result illustrated that the PPE can largely affect health risks associated with bioaerosol (Haas et al., 2017; Carducci et al., 2018) . Therefore, workers exposed to the mechanical aeration modes are strongly suggested to wear PPE. However, the effects of employing PPE on the M aeration tank in plants A and B showed weaker impact on the variability of the health risks.

This result disclosed that the microporous aeration mode did not exert obvious effects on the health risks of the workers wearing PPE as large as that on the mechanical aeration modes. This finding is consistent with previous studies that QMRA could be used to indicate the most suitable scenario to employ PPE by considering its efficiency of protection (Carducci et al., 2018) . In addition, the effective use of PPE can significantly decrease the worker's health risks (Ikehata, 2013; Haas et al., 2017) .

The P (a)inf of the workers equipped without PPE exposed to S. aureus or E. coli bioaerosols in the two WWTPs considerably exceeded the U.S. EPA benchmark workers (without or with PPE) exposed to S. aureus or E. coli bioaerosols in various aeration tanks of the two wastewater treatment plants referring to (a) workers without PPE exposed to S. aureus bioaerosol in wastewater treatment plant A, (b) workers without PPE exposed to S. aureus bioaerosol in wastewater treatment plant B, (c) workers with PPE exposed to S. aureus bioaerosol in wastewater treatment plant A, (d) workers with PPE exposed to S. aureus bioaerosol in wastewater treatment plant B, (e) workers without PPE exposed to E. coli bioaerosol in wastewater treatment plant A, (f) workers without PPE exposed to E. coli bioaerosol in wastewater treatment plant B, (g) workers with PPE exposed to E. coli bioaerosol in wastewater treatment plant A, and (h) workers with PPE exposed to E. coli bioaerosol in wastewater treatment plant B.

Correlation coefficient values were obtained from @ Oracle Crystal Ball sensitivity analyses and are shown next to each bar. for workers (without or with PPE) exposed to S. aureus or E. coli bioaerosols in various aeration tanks of the two wastewater treatment plants referring to (a) workers without PPE exposed to S. aureus bioaerosol in wastewater treatment plant A, (b) workers without PPE exposed to S. aureus bioaerosol in wastewater treatment plant B, (c) workers with PPE exposed to S. aureus bioaerosol in wastewater treatment plant A, (d) workers with PPE exposed to S. aureus bioaerosol in wastewater treatment plant B, (e) workers without PPE exposed to E. coli bioaerosol in wastewater treatment plant A, (f) workers without PPE exposed to E. coli bioaerosol in wastewater treatment plant B, (g) workers with PPE exposed to E. coli bioaerosol in wastewater treatment plant A, and (h) workers with PPE exposed to E. coli bioaerosol in wastewater treatment plant B.

Contribution to variance values were obtained from @ Oracle Crystal Ball sensitivity analyses and are shown next to each pie. J o u r n a l P r e -p r o o f Journal Pre-proof >Even without PPE, DB of workers exposed to E. coli bioaerosol was still acceptable>The use of PPE effectively reduced the health risks to an acceptable level>With high exposure concentration, input exposure parameters highly impact health risk>Mechanical aeration modes' contribution to health risks' variability was dominated>Health risks were highly influenced by the various choices of the dose- 

Application of the Crystal Ball Software for Uncertainty and Sensitivity Analyses for Predicted Concentration and Risk Levels

Evaluation of the environmental impact of microbial aerosols generated by wastewater treatment plants utilizing different aeration systems

Quantitative Microbial Risk Analysis for Various Bacterial Exposure Scenarios Involving Greywater Reuse for Irrigation

Quantitative Microbial Risk Assessment for Workers Exposed to Bioaerosol in

Wastewater Treatment Plants Aimed at the Choice and Setup of Safety Measures

Microbiology of Aerosols

DALY calculation in practice: a stepwise approach

Estimation of Infectious Risks in Residential Populations Exposed to Airborne Pathogens During Center Pivot Irrigation of Dairy Wastewaters

The application of quantitative microbial risk assessment to natural recreational waters: A review

Disease burden due to gastrointestinal pathogens in a wastewater system in Kampala

Airborne Microorganisms Emitted from Wastewater Treatment Plant Treating Domestic Wastewater and Meat Processing Industry Wastes

Microbial Dose Response Modeling: Past, Present, and Future

Quantitative Microbial Risk Assessment

Risks from Ebolavirus Discharge from Hospitals to Sewer Workers

Quantitative Microbial Risk Assessment Models for Consumption of Raw Vegetables Irrigated with Reclaimed Water

Use of floating balls for reducing bacterial aerosol emissions from aeration in wastewater treatment processes

Wastewater Reuse and Management

Quantitative microbial risk assessment of bioaerosols from a manure application site

Bioaerosol Deposition to Food Crops near Manure Application: Quantitative Microbial Risk Assessment

Relative contributions of transmission routes for COVID-19

among healthcare personnel providing patient care

Particle size distribution of cultivable airborne microbes and inhalable particulate matter in a wastewater treatment plant facility

Characteristics of airborne bacteria and fungi in some Polish wastewater treatment plants

Assessment of public health risk associated with viral contamination in harvested urban stormwater for domestic applications

Reevaluation of health risk benchmark for sustainable water practice through risk analysis of rooftop-harvested rainwater

Improved impact assessment of odorous compounds from landfills using Monte Carlo simulation

Exposure factors hankbook of Chinese population

Influence of seasons and sampling strategy on assessment of bioaerosols in sewage treatment plants in Switzerland

A role for environmental engineering and science in preventing bioaerosol-related disease

Screening-Level Microbial Risk Assessment of Urban Water Locations: A Tool for Prioritization

Quantitative microbial risk assessment of Greywater on-site reuse

Hygienic sustainability of site location of wastewater treatment plants

Seasonal changes in the concentrations of airborne bacteria emitted from a large wastewater treatment plant

Assessment of indoor airborne contamination in a wastewater treatment plant

A Systematic Review of Beef Meat Quantitative Microbial Risk Assessment Models

Occurrence and exposure assessment for the final Long Term 2

United States Environmental Protection Agency