Distribution and human health risk of N-nitrosamines in tap water in the central region of South Korea
Article information
Abstract
This study aims to evaluate the concentrations and potential risks associated with seven volatile N-nitrosamines (NAs), a category of disinfection by-products, present in tap water, while considering realistic human exposure. Tap water samples were collected across four seasons from three central regions (Gangwon State, Gyeonggi-do, and Seoul Metropolitan City) in South Korea and analyzed for the NAs using high-performance liquid chromatography with fluorescence detection, following solid-phase extraction and derivatization. Among the NAs examined, three compounds, namely N-nitrosomorpholine (NMOR), N-nitrosodimethylamine (NDMA), and N-nitrosodiethylamine (NDEA), were identified in all samples. NDMA had the highest concentration at 53.4 ± 19.5 ng/L, while NMOR and NDEA had 3.83 ± 10.56 ng/L and 37.5 ± 25.6 ng/L, respectively. These compounds exhibited similar concentrations across the three regions, with higher levels observed during winter than other seasons. The estimated cancer risk of NDEA was above 10-6, which is a concern. However, actual risk might be even lower when considering Korean drinking water intake patterns such as hot soups and stews. Nevertheless, it is imperative to improve regulatory practices to comprehensively address health risks from food intake, air contamination, and other sources.
Introduction
N-Nitrosamines (NAs) are carcinogenic compounds that can form unintentionally during water disinfection processes involving chlorination or chloramination. These harmful substances can be found in treated water, including drinking water, wastewater, and swimming pool water [1–3]. Various regulatory agencies, such as the U.S. Environmental Protection Agency (U.S. EPA), the World Health Organization (WHO), and health authorities in countries like Japan and Canada, have implemented measures for monitoring and controlling NAs, with a particular focus on N-nitrosodimethylamine (NDMA). NDMA is not only a frequently detected compound in chlorinated water, but it also poses a higher oral cancer risk (7 ng/L at a risk level of 10-5) compared to other NAs (20–60 ng/L at a risk level of 10-5 for N-nitrosodibutylamine (NDBA), N-nitrosopyrrolidine (NPYR), N-nitrosodiphenylamine (NDPA), and N-nitrosomethylethylamine (NMEA) [2,9,21,22,34,57-60]. The Office of Environmental Health Hazard Assessment (OEHHA) in the U.S. has established a public health target for NDMA in drinking water at 3 ng/L [4–7]. Health Canada has set a maximum acceptable concentration (MAC) of 40 ng/L for NDMA, while the WHO has set a guideline value of 100 ng/L. The U.S. EPA has also provided NDMA guidance values for drinking water and groundwater, ranging from 0.69 ng/L to 18 ng/L across 15 states [6]. In South Korea, NDMA and N-nitrosodiethylamine (NDEA) have been designated as key compounds for water quality monitoring since 2018 due to their carcinogenic nature [8].
NDMA poses a carcinogenic hazard, particularly associated with bladder cancer resulting from the consumption of chlorinated water. Various epidemiological studies have established a correlation between NDMA exposure and the development of different types of cancer, such as cancers affecting the upper digestive tract, brain, and stomach [2]. Despite being predominantly acknowledged for its carcinogenic attributes, NDMA is also suspected of inducing hepatotoxic effects, potentially leading to liver cirrhosis and impacts on the immune system [4]. Furthermore, NDEA is considered to be more potent than NDMA, with a cancer risk estimated at 10-5 when present at a concentration of 2 ng/L, in contrast to NDMA's cancer risk of 10-5 at 7 ng/L [9,10]. Another compound, N-nitrosomorpholine (NMOR), is recognized for its carcinogenic toxicity, affecting organs such as the liver, lungs, nasal cavity, and kidneys [11]. In toxicological studies, a metabolic mechanism has been proposed in which NAs undergo α-hydroxylation to form α-hydroxy N-nitrosamines, which then decompose into products such as aldehydes, alkyl diazohydroxides, and alkyl diazonium ions. These studies, conducted in vitro using rat liver tissue subfractions and isolated hepatocytes, suggest that the alkyl diazonium ion is a highly reactive alkylating agent that can bind to electrophilic DNA, RNA, and proteins, potentially leading to cancer. In vivo studies involving liver and kidney DNA in rats have shown that O6-methylguanine DNA adducts, associated with tumor formation, result from NDMA absorption through oral exposure in various species, including rats, mice, and hamsters. DNA alkylation following α-hydroxylation and the formation of alkyl diazonium ions is considered the main metabolic pathway for NDMA and other NAs, such as NDEA and NMOR [4,46].
The formation of NAs during water treatment processes, such as chloramination, ozonation, and chlorination, is believed to result from secondary aliphatic amines (SAAs) [12,13]. SAAs, such as dimethylamine and diethylamine, can occur naturally through metabolic processes in microorganisms, plants, and animals, or can be synthetically produced from industrial activities related to pharmaceuticals, polymers, pesticides, and other substances [14,15]. The proposed pathway for NDMA formation involves a nucleophilic substitution reaction between SAAs and monochloramine, resulting in unsymmetrical dimethylhydrazine (UDMH), which is then oxidized under chlorinated conditions. This pathway has been revised to account for reaction rates: chlorinated UDMH (Cl-UDMH) is formed from dimethylamine reacting with dichloramine and then converted to NDMA by oxygen or chloramines [3]. Some treatment plants have introduced advanced treatment processes, such as ozonation and granular activated carbon [35]. However, nitrosation of secondary amines by ozonation or chlorination of nitrite, as well as the catalytic formation on the surface of activated carbon after the fixation of reactive nitrogen species from the air, can lead to the occurrence of NAs in drinking water [12,13]. In South Korea, chlorination is the primary method used for chemical treatment of drinking water, with residual chlorine levels required to be above 0.2 mg/L to prevent microbial contamination [16]. The formation of NDMA through chlorination has been associated with the interaction between free chlorine and ammonia, leading to chloramine formation [17–19]. Besides SAAs, chloramines, and chlorine, other factors influencing NA formation include dissolved oxygen, natural organic matter, and additional constituents [3,20].
NAs are widely distributed in drinking water globally and have been the subject of extensive research. In Japan, NDMA levels in drinking water ranged from not detected (ND) to 2.2 ng/L in summer and ND to 10 ng/L in winter [21]. In China, nine NAs were measured in three drinking water treatment plants, from source water to tap water, with NDMA concentrations ranging from 1.6 to 9.4 ng/L in source water and 1.2 to 8.9 ng/L in tap water [22]. Other countries have also reported the presence of NDMA and other NAs in drinking water. NDMA was detected at concentrations up to 59 ng/L in chloraminated water, with higher levels observed during spring runoff due to the inflow of precursors. Additionally, NMOR, NDPA, and NPYR were detected at levels up to 6.0, 1.0, and 4.4 ng/L, respectively [23]. NDEA and NDBA were found in treated water at concentrations below 1.5 ng/L [24]. In South Korea, six NAs, including NDMA, NMEA, NDEA, NDPA, NMOR, and NDBA, were monitored in finished water at treatment plants. Among them, NDMA and NMOR had the highest detection rates of 9.22% and 17.0%, with concentrations at ND–6.7 ng/L and ND–9.5 ng/L, respectively [5]. However, monitoring finished water may not reflect potential increases in the levels of NAs due to contact time. The maximum concentration of NDMA is seven times lower than the regulatory level established by the Ministry of Environment in Korea [8]; however, this compound was detected at 47.7 ng/L in tap water in 2009 [61]. Qi et al. [25] reported a rapid increase in NDMA levels in natural water after adding monochloramine at 2 mM, suggesting that the concentrations of NAs in tap water could be underestimated. NDMA and other NAs are consistently found in treated water in South Korea at levels similar to those in other countries. Inadequate removal of precursors can lead to further formation of NAs, which is a concern. Therefore, it is crucial to conduct realistic monitoring of tap water and consider regional and seasonal variations in the concentrations of NAs, which are influenced by various physicochemical factors.
In this study, seven NAs were monitored using a quantitative approach involving high-performance liquid chromatography with fluorescence detection (HPLC-FLD). Tap water samples were collected across four seasons from three distinct regions in South Korea: Gangwon State, Gyeonggi-do, and Seoul Metropolitan Government (Seoul), which were selected to represent areas with varying population densities. The collected data underwent statistical analysis to evaluate variations in the concentrations of NAs across regions and seasons. Additionally, a human cancer risk assessment was conducted considering various exposure pathways specific to the Korean population.
Materials and Methods
Reagents and materials
All standard solutions of NAs, N-nitrosomethylbutylamine (NMBA), Carboxen 572, and sodium thiosulfate were obtained from Sigma-Aldrich (St. Louis, MO, USA). Stock solutions were individually prepared in methanol at a concentration of 1,000 mg/L, and NMBA was used as the surrogate standard. HPLC-grade acetone, methanol, and dichloromethane were purchased from Honeywell Burdick & Jackson Inc. (Muskegon, MI, USA). Sodium hydroxide, sodium bicarbonate, sodium sulfate, hydrochloric acid (35%), and acetic acid were obtained from Daejeong (Siheung, Korea). 5-(Dimethylamino)naphthalene-1-sulfonyl chloride (dansyl chloride) was purchased from Calbiochem (San Diego, CA, USA), and hydrobromic acid (48%) was procured from Wako (Osaka, Japan).
A denitrosation solution was prepared by diluting 1 mL of 48 % hydrobromic acid with acetic acid to a final volume of 10 mL. A dansylation solution was prepared by dissolving 25 mg of dansyl chloride in 50 mL of acetone. A 50 mL buffer solution with a pH of 10.5 was prepared by dissolving 0.6 g of sodium hydroxide and 2.0 g of sodium bicarbonate in ultrapure water. All solutions were stored at 4 °C and used within 2 weeks.
Reagents and materials Field sampling
Tap water samples were collected across four seasons: winter (January 28 to February 1, 2013), spring (May 8 to May 22, 2013), summer (August 1 to August 15, 2013), and fall (October 15 to October 29, 2013). Sampling locations are detailed in Fig. 1, including five sites in Gangwon State, twelve sites in Gyeonggi-do, and thirteen sites in Seoul. All samples were collected in 1 L dark brown glass bottles without headspace from the taps in community centers. Approximately 0.2 g of sodium thiosulfate was added to each sample for dichlorination, following U.S. EPA Method 521 [28]. The samples were transported to the laboratory in a cooler and stored at 4 °C until analysis. On site, a portable device (MARTINI Mi411, Milwaukee Instruments, Inc., Rocky Mount, USA) was used to measure residual chlorine and pH.
Map of South Korea (A) and the sampling regions (B): Gangwon (five green square), Gyeonggi (twelve red triangle), and Seoul (thirteen blue circle) (B).
*Tap water in Gyeonggi and Seoul is sourced from the Han River, while in the Gangwon region, some sampling areas use local water sources.
Quantitative determination of N-nitrosamines in tap water
The analytical method for NAs used in this study was an optimized version based on Jung et al. [26], which was a modification of the U.S. EPA Method 521 and involved derivatization using dansyl chloride [27,28]. NAs can be identified and quantified after solid-phase extraction with activated coconut charcoal and gas chromatography-tandem mass spectrometry, as described in U.S. EPA Method 521. An additional derivatization process allows for quantification at ng/L levels using HPLC-FLD, achieving high accuracy and precision. Each 500 mL sample was spiked with 1 µ L of NMBA (20 mg/L, used as a surrogate standard). The sample was then passed through a cartridge containing 2.0 g of Carboxen 572 at a flow rate of 10 mL/min. The cartridge was connected to an impinger with 16 g of silica gel and dried using a vacuum pump (at -34 kPa) for 60 min. The NAs were eluted from the cartridge using 15 mL of dichloromethane at a flow rate of 4 mL/min. The eluate was then concentrated to 1 mL under a gentle stream of nitrogen.
After adding 150 µ L of the denitrosation solution, the concentrate was vigorously mixed for 10 s. The mixture was then incubated at 40 °C for 30 min and subsequently dried under a gentle stream of nitrogen. Next, 150 µ L of a buffer solution (pH 10.5) and 150 µ L of the dansylation solution were added to the sample and agitated vigorously for 10 s. The sample was incubated at 40 °C for an additional 30 min, followed by the addition of 50 µ L of ultrapure water. Finally, the analytcal sample was transferred to a 1 mL vial, with a 30 µ L injection volume.
For HPLC-FLD analysis, a Varian Inc. 410 autosampler (Palo Alto, USA) was used for sample injection. The column used was a Microsorb-MV 100-5 C18 (5 µ m, 250 mm × 4.6 mm). The pumps used were Varian Inc. Prostar 210, and the mobile phase consisted of water and acetonitrile in a 45:55 (v/v) ratio, with a flow rate of 1 mL/min. Detection was carried out using a Waters 747 scanning fluorescence detector (Milford, USA), with excitation and emission wavelengths set at 340 nm and 530 nm, respectively.
The method validation results are summarized in Table 1. Calibration curves were established at five concentration levels ranging from 2 ng/L to 80 ng/L, with linearities (r2) exceeding 0.99. Accuracy and precision were evaluated in terms of recovery and relative standard deviation (RSD). The average recoveries (n = 3) ranged from 95.7% to 99.3%, while the RSDs (n = 6) ranged from 4.37% to 15.2%. The detection limit was determined based on a signal-to-noise ratio of 3. The method detection limits (MDL) ranged from 0.92 to 1.45 ng/L, and the limits of quantification (LOQ) ranged from 3.35 to 5.25 ng/L.
Method validation for seven N-nitrosamines in water.
Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics 29 software at a 5 % of significance level. A one-way analysis of variance (ANOVA) was conducted to evaluate variations in concentration distributions, followed by post-hoc analysis using Tukey’s HSD test to identify specific variations between groups. The normal distribution of the data was tested using the Kolmogorov-Smirnov and Shapiro-Wilk tests. Data sets that were not normally distributed were analyzed using nonparametric tests, including the Mann-Whitney U test, Kruskal-Wallis test, and Spearman's rank correlation analysis.
Risk assessment
Ingestion
According to the Risk Assessment Guidance for Superfund Volume I: Human Health Evaluation Manual, Volume III - Part A [29,30], the ingestion risk of chemicals in drinking water is calculated using specific equations. The exposure dose was calculated for both central tendency exposure (CTE) and reasonable maximum exposure (RME) scenarios, using the following equation:
Where LADD is the lifetime average daily dose (mg/kg-day), Cw is the concentration in water (mg/L), IR is the ingestion rate (L/day), EF is the exposure frequency (days/year), ED is the exposure duration (years), BW is the body weight (kg), and AT is the averaging time. The AT is calculated by multiplying the Korean average life expectancy by 365 days/year. Excess lifetime cancer risk (ELCR) is calculated using the following equations.
Where, SFO is the oral cancer slope factor (mg/kg-day)-1.
Inhalation
According to the Risk Assessment Guidance for Superfund Volume I: Human Health Evaluation Manual, Part A [30,31], the inhalation risk of airborne chemicals (vapor phase) in drinking water is calculated using the following equations.
Where CA is the concentration in air (mg/m3), calculated using the above equation. H is Henry’s law constant (atm-m3/mol at 37 °C), R is the gas constant (0.0821 L-atm/mol-K), and T is the temperature (37 °C). InhR is the inhalation rate (m3/h), ET is the exposure time (h/day), EF is the exposure frequency (days/year), ED is the exposure duration (years), BW is body weight (kg), and AT is the averaging time. The AT is calculated by multiplying ED by 365 days/year. ELCR is calculated using the following equations.
Where IUR is the inhalation unit risk (μg/m3)-1, with values of 0.0019 for NMOR, 0.014 for NDMA, and 0.043 for NDEA. The exposure concentration (EC, mg/m3) is calculated as follows: EC=((CA×ET×EF×ED))⁄AT.
Dermal
According to the Risk Assessment Guidance for Superfund Volume I: Human Health Evaluation Manual, Part E [32], the dermal absorbed dose per event for organic compounds and the dermal absorbed dose from water contact are calculated using the following equations.
Where t* is the time to reach steady-state (h), tevent is the event duration (h/event), DAevent is the absorbed dose per event (mg/cm2-event), FA is the fraction of water absorbed (unitless), Kp is the dermal permeability coefficient of the compounds in water (cm/h), Cw is the concentration in water (mg/cm3), and τevent is the lag time per event (h/event). The values for t*, FA, Kp, and τevent are specific to each chemical. For NMOR, NDMA, and NDEA, these values are as follows: t*= 1.14, 0.67, and 0.80 h; FA = 1.0 for all three; Kp = 0.00018, 0.00025, and 0.001 cm/h; and τevent = 0.48, 0.28, and 0.33 h, respectively.
LADD and ELCR are calculated using the following equations.
Where EV is the event frequency (events/day), ED is the exposure duration (years), EF is the exposure frequency (days/year), SA is the skin surface area available for contact (cm2), BW is the body weight (kg), and AT is the averaging time (days).
Where SF is the absorbed cancer slope factor (mg/kg-day)-1, calculated as follows: SFABS=SFO⁄ABSGI . Here, SFO is the oral slope factor (mg/kg-day)-1, with values provided for NMOR, NDMA, and NDEA. ABSGI is the fraction of the contaminant absorbed in the gastrointestinal tract (unitless), assumed to be 1.0 for 100% absorption.
Results and Discussion
Determination of N-nitrosamines in the central region of South Korea
Seven NAs were analyzed in a total of 120 tap water samples collected from the central region of South Korea over four seasons. The detection numbers and concentrations are detailed in Tables 2 and 3. Among the seven NAs, three compounds—NMOR, NDMA, and NDEA—were found in all samples. In contrast, NMEA, NPYR, NPIP, and NDPA were detected in only 2, 3, 17, and 13 out of 120 samples, respectively, indicating low detection rates (1.7–14%). The sum of the seven NAs averaged 95.7 ng/L, with concentrations range of 1.85–231 ng/L. For individual NAs, the average concentrations and ranges are as follows: NMOR had an average of 3.83 ng/L with a range of 0.550–87.8 ng/L, NDMA had an average of 53.4 ng/L with a range of 0.460–99.8 ng/L, and NDEA had an average of 37.5 ng/L with a range of 0.750–106 ng/L. NDMA showed the highest average concentration, followed by NDEA and NMOR. These compounds are classified as Group 2A or 2B carcinogens by the WHO, with oral cancer slope factors of 6.7, 51, and 150 (mg/kg-day)-1 for NMOR, NDMA, and NDEA, respectively [9–11,33]. NAs are known to be compounds that can alkylate DNA and lead to cancer. Oral exposure to NDMA and NDEA through drinking water has been shown to cause tumors in the liver, gastrointestinal tract, urinary tract, esophagus, and lungs of rats, mice, and hamsters [4,46]. Particularly, NDMA and NDEA are considered potent carcinogens, and regulations for these substances in drinking water have been enforced in South Korea since 2018. In this study, the maximum concentrations of NDMA and NDEA were 1.4 and 5.3 times higher, respectively, than the regulatory limits in South Korea, which are 70 ng/L for NDMA and 20 ng/L for NDEA [8].
Regional distribution of seven N-nitrosamines (NAs) in tap water samples
Seasonal distribution of seven N-nitrosamines (NAs) in tap water samples.
Among the samples, 20, 48, and 52 samples were collected from Gangwon, Gyeonggi, and Seoul, respectively. The results of the regional distribution are shown in Table 2 and Fig. 2. The sum of the seven NAs was detected at 106 ng/L in Gangwon, 90.7 ng/L in Gyeonggi, and 96.5 ng/L in Seoul, with target NAs at similar levels across the three regions. For individual NAs, NMOR was found at an average of 2.88 ng/L in Gangwon, 3.80 ng/L in Gyeonggi, and 4.22 ng/L in Seoul, with no significant difference observed (p > 0.05). However, NMOR was detected up to 87.8 ng/L in Seoul, approximately four times and two times higher than the maximum concentrations found in Gangwon and Gyeonggi, respectively. Previous studies have reported that morpholine, the precursor of NMOR, was found at 0.27 μg/L in only one of 82 drinking water samples in South Korea. NMOR was detected up to 9.5 ng/L with a 17.0% detection rate (35/206), showing higher maximum concentrations and detection rates compared to other NAs (NDMA, NMEA, NDEA, NDPA, and NDBA) [5,34]. Despite differences in specific treatments at drinking water treatment plants in Gyeonggi and Seoul, such as advanced water treatment using granular activated carbon after ozonation, both regions source their water from the Han River [35]. There is a possibility that NMOR may exist as a contaminant in the source water. Previous studies have shown that NMOR was found in most influent and effluent samples at plants and did not increase during treatment, which could be due to groundwater contamination issues related to nitrate [23]. Morpholine is not commonly found in nature but is anthropogenically used as an intermediate in rubber processing, a corrosion inhibitor in steam condensation in the gas and pipeline industries, and in automobiles, leather, and furniture products such as waxes and polishes. Additionally, it has been used in food packaging and has been detected in food and food packaging materials at concentrations ranging from ND to 0.84 mg/kg [47-50]. Seoul, the capital city of South Korea, has a much higher population density of 15,333 people/km2, compared to 1,351 people/km2 in Gyeonggi and 91 people/km2 in Gangwon [51]. Therefore, NMOR could enter source water through human activities, such as domestic wastewater, in densely populated areas like Seoul. Further investigation is needed into factors related to human and agricultural activities, such as natural organic matter (NOM), dissolved organic nitrogen (DON), and tertiary alkylamines [36,37]. Despite the limitations due to the lack of data and low detection rates, risk assessments should be conducted to verify the need for monitoring and regulation.
Box plots of N-nitrosomorpholine (NMOR), N-nitrosodimethylamine (NDMA), N-nitrosodiethylamine (NDEA), and sum of N-nitrosamines in tap water collected from three regions (Gangwon, Gyeonggi, and Seoul) in South Korea by four seasons.
A significant difference between regions was confirmed using an analysis of variance (ANOVA) test at the following significance levels: *p < 0.05, **p < 0.01, and ***p < 0.001.
NDMA and NDEA were found at concentrations ranging from 0.460 to 99.8 ng/L and 0.750 to 106 ng/L, respectively, with similar concentrations across the three regions (p > 0.05). However, NDEA concentrations in summer showed significant regional variation, with Gangwon exhibiting levels of 27.0 ng/L—three times and two times higher than those in Gyeonggi and Seoul, respectively (p < 0.01 and p < 0.05). Although NDMA concentrations did not show significant regional differences, the average concentration in Gangwon was twice as high as in Gyeonggi. Diethylamine is used in various applications, including as an additive in fuels and oils, an epoxy hardener, flotation agents, pharmaceuticals, corrosion inhibitors, and pesticides [52]. The highest concentrations in Gangwon were 75.9 ng/L in Sokcho and 62.3 ng/L in Chuncheon. The region contains established agricultural and industrial complexes, including farming areas as well as pharmaceutical, cosmetic, and food industries. During industrial activities and rainfall periods, there is a possibility that diethylamine could enter the source water due to soil runoff.
Several factors could contribute to these differences, including the presence of NDMA, its precursors, and other influencing factors. NDMA can be released into the environment through wastewater from chemical plants. In the U.S., effluent from wastewater treatment plants using ozonation has been reported to contain NDMA at concentrations range of 1–537 ng/L [38]. Dimethylamine, a precursor of NDMA, was found at 0.92 μg/L in Gangwon in a previous study, higher than the 0.81 and 0.57 μg/L found in Gyeonggi and Seoul, respectively [39]. The Gangwon area, which is 80% forested, may have a higher background concentration of precursors due to natural occurrences by organisms and agricultural activities such as pesticide use. Additionally, effluent might influence NDMA concentrations during summer due to concentrated precipitation. To clarify these issues further, additional studies are needed to investigate NAs, their precursors, and other physicochemical factors in both source and treated water.
Seasonal distribution of N-nitrosamines in tap water
Seven NAs were seasonally observed in tap water, as described in Table 3 and Fig. 3. For the three NAs with high detection rates, the annual average concentrations are 3.83 ng/L for NMOR, 53.4 ng/L for NDMA, and 37.5 ng/L for NDEA. The sum of the seven NAs was detected at an average of 95.7 ng/L. Seasonal average concentrations of these three NAs are as follows: NMOR was 10.3 ng/L in winter, 1.67 ng/L in spring, 0.979 ng/L in summer, and 2.37 ng/L in fall; NDMA was 64.0 ng/L in winter, 53.2 ng/L in spring, 51.6 ng/L in summer, and 44.7 ng/L in fall; NDEA was 54.8 ng/L in winter, 40.1 ng/L in spring, 27.0 ng/L in summer, and 28.2 ng/L in fall. The sum of NAs was highest in winter, followed by spring, summer, and fall. Seasonal distributions showed that NMOR was highest in winter, followed by fall, spring, and summer; NDMA was highest in winter, followed by spring, summer, and fall; and NDEA was highest in winter, followed by spring, fall, and summer. Generally, the concentrations were highest in winter. This result is consistent with previous studies that reported higher concentrations of NAs in source and tap water during dry seasons compared to heavy rainfall periods [22].
Box plots of N-nitrosomorpholine (NMOR), N-nitrosodimethylamine (NDMA), N-nitrosodiethylamine (NDEA), and sum of N-nitrosamines in tap water collected in four seasons in South Korea by three regions (Gangwon, Gyeonggi, and Seoul).
A significant difference between regions was confirmed using an analysis of variance (ANOVA) test at the following significance levels: *p < 0.05, **p < 0.01, and ***p < 0.001.
To confirm significant differences between the four seasons, ANOVA was performed, and the results are marked as * in Fig. 3. The sum of the seven NAs is significantly higher in winter compared to other seasons (p < 0.01), with the highest concentrations observed in Gyeonggi and Seoul. Gangwon samples, however, showed similar levels in winter and summer, with no significant seasonal difference (p > 0.05). For individual compounds, NMOR concentrations are higher in winter than in other seasons (p < 0.05), with the highest levels observed in Gyeonggi during winter, range of 0.550–49.4 ng/L. NDMA and NDEA in winter also showed significantly higher concentrations compared than other seasons (p < 0.05). NDMA concentrations in Seoul were significantly higher in winter than in fall, the lowest season. Conversely, NDMA in summer was highest in Gangwon, with the widest distribution observed in Gyeonggi. Similar patterns were also observed for NDEA. Previous studies of tap water showed similar results, with dimethylamine concentrations in Seoul being slightly higher in winter than in summer, and a negative correlation between NDMA and temperature observed in tap water [34,39]. In contrast, tap water in Gangwon showed dimethylamine levels twice as high in summer as in winter [39]. Dimethylamine is widely used in various industries, including rubber, detergents, pesticides, and pharmaceuticals. Both secondary amines and NAs are present in foods such as beer, fish, vegetables, and dairy products, and can be excreted through human physiological activity [46,53]. Therefore, NDMA and its precursors can readily enter the water environment. NDMA was found at concentrations ranging from 15.7 to 2320 ng/L in river water, both upstream and downstream of sewage treatment plants [62]. Their seasonal distribution may be influenced by degradation through photolysis and microorganisms [13,54,55,56]. A previous study in China also found similar results, with higher NDMA concentrations in treated water and a 1.7-fold higher removal rate in the sedimentation process during summer compared to winter. On the other hand, the higher concentrations in summer in Gangwon are suspected to result from temporary runoff caused by heavy precipitation in nearby agricultural lands, as well as pharmaceutical, food, and cosmetic industries. Wastewater may have different water quality, containing more hydrophilic compounds (e.g., DON) than hydrophobic compounds (e.g., humic and fulvic acids), which could affect NDMA formation potential [25]. Unfortunately, the small sample size from Gangwon and the lack of clear differences in weather conditions, such as precipitation and temperature, between Gangwon and other regions, limit the ability to draw definitive conclusions.
For the three most frequently detected NAs, Spearman correlation analysis was performed (Fig. 4). Although no significant correlation with NAs was found for residual chlorine and pH (p > 0.05), a strong positive correlation was observed between NDMA and NDEA (p < 0.001). The occurrence of these compounds might be influenced by similar sources or suggest the possibility of their formation within the water treatment systems. Despite the lack of significance in the present study, previous research has reported higher NDMA formation in chlorinated water at pH 6 compared to pH 7 and 8 [40]. However, identifying a clear correlation between chlorination and NAs in tap water remains challenging. Krasner et al. [23] also noted the difficulty in pinpointing seasonal characteristics due to regional, temporal, and geographical variations. Additionally, NDMA formation potential has shown correlation (R2) values of 0.60 and 0.78 with dissolved organic carbon and nitrogen, respectively [25]. Given the complexity of NA formation due to numerous physicochemical factors, further studies should incorporate additional factors, as mentioned above, to gain a clearer understanding.
Correlation plot of N-nitrosamines and factors in tap water collected from central regions in South Korea (Each value within the graph represents a correlation coefficient).
A significant difference between regions was confirmed using a Spearman's rank correlation analysis at the following significance levels: *p < 0.05, **p < 0.01, and ***p < 0.001.
Human cancer risk
Human cancer risk assessment was conducted for NMOR, NDMA, and NDEA, with results detailed in Table 4. For ingestion, the ELCRs are as follows: NMOR shows ELCRs of 4.03 × 10-7 for CTE and 9.23 × 10-6 for RME; NDMA has ELCRs of 3.06 × 10-6 for CTE and 8.22 × 10-5 for RME; and NDEA presents ELCRs of 8.83 × 10-5 for CTE and 2.49 × 10-4 for RME. According to the WHO, the acceptable risk range is between 10-6 and 10-4. NDEA exceeds the upper limit of this range, with ELCRs of 2.57 × 10-4 for males and 2.42 × 10-4 for females in the RME scenario. The ELCRs for NDMA fall within the acceptable range but are close to the upper limit at 8.22 × 10-5. For other exposure routes, ELCRs range from 7.30 × 10-17 to 1.36 × 10-11 for inhalation and from 1.22 × 10-9 to 5.08 × 10-6 for dermal exposure. These values are significantly lower than the upper acceptable risk limit of 10-4, indicating that both inhalation and dermal exposure routes are not expected to have adverse effects on human health.
Human carcinogen risk assessment of three N-nitrosamines in tap water
In the present study, ingestion of drinking water is identified as the primary exposure route, with NDMA and NDEA raising concerns due to their carcinogenicity. Despite various consumption methods in South Korea—such as drinking boiled water, using water purifiers, or consuming bottled water—additional exposure sources, including smoking, food, and beverages, also contribute to overall risk. The WHO has set a guideline value for NDMA levels in drinking water at 100 ng/L [7]. The Ministry of Environment in South Korea has established monitoring levels for NDEA and NDMA at 70 ng/L and 20 ng/L, respectively, in drinking water [8]. These levels correspond to an oral risk level of 10-4, as estimated by the U.S. EPA [9,10]. In our study, the average concentration of NDEA (37.5 ng/L) exceeds the monitoring level, and its risk (8.83 × 10-5 for CTE) is close to the 10-4 threshold. Furthermore, the NDMA concentration is also near the cancer risk threshold of 10-4 for RME, which is not negligible when considering other exposure routes, such as food and smoking. NDMA, NDEA, and NMOR have been detected in agricultural, livestock, and aquatic products at concentrations of up to 10.4 μg/kg, 284 μg/kg, and 14.0 μg/kg, respectively. Notably, NDMA and NDEA levels increase after cooking processes, such as air-frying and boiling [41]. Additionally, naturally occurring phenols in food can act as catalysts for the formation of NAs from nitrates and secondary amines [42]. In tobacco, NDMA concentrations reach up to 39.8 ng/g in unburned tobacco and 7.9 ng per cigarette in mainstream smoke [43].
To protect public health and ensure the provision of clean drinking water, it is crucial to reduce NDMA and NMOR levels to significantly lower than those indicated by this study's risk assessment. Various methods for removing NDMA have been reported, including the use of carbon filters and ultrafiltration membrane units [22], as well as granular activated carbon (GAC) following ozonation [24]. Breakpoint chlorination or ammonia removal are alternative strategies to mitigate NA formation [7,44,45]. More than 70% of NDMA precursors in wastewater were removed by biological treatment, and removal rates greater than 90% were observed when using a membrane bioreactor coupled with nanofiltration [63,64]. In previous studies, zeolites and powdered activated carbon (PAC) were tested for precursor removal, and it was suggested that a combination of adsorbents be considered based on the removal rates of amines, depending on their polarity [65]. The lowest levels of NAs were observed in a nationwide monitoring study of drinking water treatment plants in South Korea when using a combination of filtration, ozonation, and granular activated carbon (GAC) treatment [66]. Future research should focus on developing cost-effective removal processes or alternative disinfection strategies to address this issue effectively.
Conclusions
This study investigated the seasonal distribution of seven NAs in tap water across three regions in South Korea (Gangwon, Gyeonggi, and Seoul). This approach provides a more realistic assessment of human exposure compared to monitoring systems in drinking water treatment plants. Among the NAs, NMOR, NDMA, and NDEA were the primary compounds detected, with concentrations ranging from 0.460 ng/L to 99.8 ng/L. Although the distribution of these NAs was similar across the sampling regions, potential influences from agricultural, industrial, and anthropogenic activities remain a concern. NAs were generally higher in winter compared to other seasons, but there were limitations in understanding their occurrence due to small sample sizes and insufficient data on precursors and other influencing factors. The study included carcinogenic risk assessments via various exposure routes (ingestion, inhalation, and dermal). Although the risk levels were below 10-4, reductions are still needed when considering other exposure sources such as food and tobacco. Future efforts should focus on monitoring disinfection byproducts like NAs in tap water more closely and strengthening regulations to safeguard human health.
Notes
Conflict of interest
All authors declare that they have no conflicts of interest.
CRediT author statement
DP: Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing; SJ: Data curation, Formal analysis, Investigation, Methodology, Validation; DK: Data curation, Formal analysis, Investigation, Methodology, Validation; HK: Conceptualization, Project administration, Supervision, Writing – review & editing.