Presence of Disinfection Byproducts in Public Swimming Pools in Medellín, Colombia

The quality of water in swimming pools is essential to avoid risks to the health of users. Medellín has more than 1000 public swimming pools, which are supervised by the Medellín Health Authority to monitor and ensure compliance with relevant regulations. The Health Authority has financed several studies related to the quality of drinking and recreational water in Medellín in order to protect consumers and users. One such study involves the evaluation of the presence of disinfection byproducts (DBP). The best known DBPs resulting from disinfection with chlorine are trihalomethanes (THMs) and halogenated acetic acids (HAAs), as well as other minorities such as chloramines or halophenols (HPs). DBPs pose a greater risk in swimming pool water because there is a greater possibility of ingestion, since exposure occurs through several routes at the same time (direct ingestion of water, inhalation of volatile or aerosol solutes, dermal contact and absorption through skin). In the present work, high concentrations of THMs and HAAs were detected in the public swimming pools selected in the study, but the presence of HPs was not detected in the pools.


Introduction
Disinfection byproducts (DBPs) are toxic, bioaccumulative, and carcinogenic, and their formation occurs through the reaction between the disinfectant and organic compounds [1]. There are several disinfection techniques to eliminate pathogenic microorganisms, among which are ultraviolet radiation, solar disinfection, chlorine dioxide, slow filtration, and the use of chemical compounds such as hydrogen peroxide, ozone, and chlorine compounds. Chlorine compounds are the most commonly used due to their low cost, disinfecting power, residuality, biocidal properties, and easy use [2]. In swimming pool disinfection, the most commonly used in Medellín are sodium hypochlorite and calcium hypochlorite [3]. The residual free chlorine is HOCl (hypochlorous acid) and OCl − (hypochlorite), being that HOCl is a much better oxidizing and disinfecting agent than OCl − . The Secretary of Health of Medellín established through Resolution 2018060366702 of 2018, a concentration of residual free chlorine between 1.0-5.0 mg/L for pools. Residual free chlorine generates DBPs [4], which are a group of organic and inorganic compounds formed by reactions between disinfectants and organic matter present in water [5].

Samplings in Selected Public Swimming Pools
The samples were taken by means of point sampling in the pools of six public establishments. One of these establishments (B) occupies 17 hectares, with several pools, facilities for the practice of different athletic disciplines, and a bike path. Another is the Medellín an aquatic complex (C), with a monthly admission of about 45,000 people, and where highly competitive swimmers from Medellín practice.
In each swimming pool, four monitoring sessions were carried out: in times of high influx of bathers, in times of low influx of bathers, on a rainy day, and on a sunny day. Sampling during periods of high and low influx of bathers was scheduled according to the holiday season or weekends.

Disinfection Byproducts (DBPs)
The samples were maintained at 0 • C while being transported in glass containers to the Pollution Diagnostic and Control Group-GDCON laboratory. THMs, HAAs, and HPs analyses were performed Int. J. Environ. Res. Public Health 2020, 17, 4659 3 of 12 using a gas chromatograph with a micro-electron capture detector (micro-ECD). For THMs, the ASTM D6520-06 [18] method was followed, which consists of the extraction of the analytes using a fused silica fiber and subsequent reading with the equipment by thermal desorption. HAAs were analyzed according to EPA method 552.2 [19], involving a liquid-liquid extraction of 40 mL of sample with 4 mL of methyl tert-butyl ether (MTBE) and analysis with a capillary gas column (GC primary column-DB-5625). The HP analyses were performed according to the UNE-EN 12,673 [20] gas chromatographic separation method.

Results of THM and HAAs Concentration
The concentrations of THMs and HAAs were higher on rainy days than on sunny days (when temperature is around 32-34 • C), which is due to the fact that the higher temperature favors the volatilization of THMs and HAAs, decreasing their concentration in the water. It is clearly observed that, when the pool has more users, the concentration of THMs is higher than when there are fewer users. This indicates that the organic matter from users contribute to forms THMs.
Most of the pools showed higher concentrations of HAAs when there were fewer users, which is surprising given the findings of Kanan and Karanfil (2011), who showed that the organic matter of the users' bodies has HAAS-forming potential [14].
Samples with few users were collected on a Saturday around 9 a.m., and samples with many users were collected at around 3 p.m. It is likely that, due to the temperature on the day in Medellín (32 to 34 • C, as the samples were collected on sunny days), the pools lost HAA concentration. The HAAs formed and accumulated until 6 p.m., the time the pools close, and perhaps did not evaporate, since from 5 p.m. the temperature of Medellín begins to decrease, reaching 19 • C between 9 p.m. and 7 a.m. Therefore, the next day at 9 a.m., the concentration of HAAs may be high. This is not the case with THMs as they are more easily volatilized, even at night.
It is important to mention that the water used to fill swimming pools is also for human consumption in the city of Medellín, and this water is of good quality (our group has been monitoring the quality of the water for human consumption in the city of Medellín from 2004 to date). The concentrations of DBPs in water for human consumption in Medellín are very low ( Table 2) and they comply with the regulations established in Colombia. All the pools selected in this study use water from the Ayurá or Manantiales aqueducts. Figures 1 and 2 show how the concentration of THMs and HAAs varied in the selected pools in Medellín according to the DBP study. The P1, P4, P5, P9, and P14 pools had the highest THMs and HAAs values. P9 corresponds to a teaching pool for babies, who generate a large amount of body fluids, favoring the production of THMs and HAAs. Eleven pools did not comply with the regulations for the maximum value of THMs in pools of the France standard and the German DIN standard ( Figure 1). Figure 3 shows that the residual free chlorine measured during sampling to measure THMs and HAAs was generally well below 2 mg/L.

Residual Free Chlorine Measured during the Monitoring of DBPs
In general, the residual chlorine influenced the concentration of THMs and HAAs. However, the speed of formation of the DBPs depends on the pH of the water in each pool. The average pH of swimming pool water was 7.55, with a maximum value of 8.45 and a minimum of 7.05. The pKa of hypochlorous acid is 7.53 (25 • C), indicating that the majority species in some pools was hypochlorite, and in others hypochlorous acid. Hansen et al. (2013) proved that formation of THMs increases in pools when pH is increased above 7.2 [21].  The P1, P4, P5, P9, and P14 pools had the highest THMs and HAAs values. P9 corresponds to a teaching pool for babies, who generate a large amount of body fluids, favoring the production of THMs and HAAs. Eleven pools did not comply with the regulations for the maximum value of THMs in pools of the France standard and the German DIN standard (Figure 1). Figure 3 shows that the residual free chlorine measured during sampling to measure THMs and HAAs was generally well below 2 mg/L. The P1, P4, P5, P9, and P14 pools had the highest THMs and HAAs values. P9 corresponds to a teaching pool for babies, who generate a large amount of body fluids, favoring the production of THMs and HAAs. Eleven pools did not comply with the regulations for the maximum value of THMs in pools of the France standard and the German DIN standard (Figure 1). Figure 3 shows that the residual free chlorine measured during sampling to measure THMs and HAAs was generally well below 2 mg/L.

Discussion
To provide a comparison reference, Table 3 shows the results of THMs and HAAs in swimming pools in other countries. Based on this, THMs were found at surprisingly high values in public pools in Medellín. The values for HAAs found in public swimming pools in Medellín were also generally the highest, but they are in the range in which these byproducts were found in the United States. Despite the fact that it is mandatory for public pools in Medellín to have foot washes and showers, users often do not comply with requirements to shower thoroughly before entering the pool, which favors the organic matter of the bodies of the users or their sun protection products, which react with the chlorinated disinfecting agent to form HAAs and THMs.  [22] and supplemented by us).  Table 4 shows the maximum allowed concentrations (MAC) of THMs in swimming pool water in several European countries. Most of the samples from the public swimming pools in Medellín have levels of THMs that would not comply with the limits shown in this table. This is very worrying, because it is putting the health of bathers at risk. The Medellín Health Authority has demanded for many years that establishments apply controls on the use of swimming pools, such as keeping them isolated from green areas, and having mandatory foot washing areas and showers. Colombia does not have reference values for DBPs in swimming pools, but for drinking water it has established a maximum allowed value of 200 μg/L in THMs; there are no reference values for the other DBPs.

Discussion
To provide a comparison reference, Table 3 shows the results of THMs and HAAs in swimming pools in other countries. Based on this, THMs were found at surprisingly high values in public pools in Medellín. The values for HAAs found in public swimming pools in Medellín were also generally the highest, but they are in the range in which these byproducts were found in the United States. Despite the fact that it is mandatory for public pools in Medellín to have foot washes and showers, users often do not comply with requirements to shower thoroughly before entering the pool, which favors the organic matter of the bodies of the users or their sun protection products, which react with the chlorinated disinfecting agent to form HAAs and THMs.  [22] and supplemented by us).  Table 4 shows the maximum allowed concentrations (MAC) of THMs in swimming pool water in several European countries. Most of the samples from the public swimming pools in Medellín have levels of THMs that would not comply with the limits shown in this table. This is very worrying, because it is putting the health of bathers at risk. The Medellín Health Authority has demanded for many years that establishments apply controls on the use of swimming pools, such as keeping them isolated from green areas, and having mandatory foot washing areas and showers. Colombia does not have reference values for DBPs in swimming pools, but for drinking water it has established a maximum allowed value of 200 µg/L in THMs; there are no reference values for the other DBPs. According to Figures 1 and 2, P19 pool is highlighted (see coding in Table 1), as it had high values of THMs and HAAs in all the samplings. This pool is used for teaching babies, who because of their age are not aware of bodily functions and involuntarily urinate inside the pool and will also have products, such as sunscreen, that their parents apply to their skin. The biggest contributor to DBPs in pools is urine; it has estimated that swimming pools contain an average of 30 to 80 mL of urine for each person [36]. Therefore, there is a high input of organic matter into this pool. Urea, the main component in urine [13], was reported in swimming pools at concentrations of 0.50-2.12 mg/L [37], 0.12-3.6 mg/L [38], 0.01-0.11 mg/L [28], and 0.23 ± 0.19 mg/L [39]. Urine has other compounds such as creatinine, hippuric acid, citric acid, ammonia, uric acid, glycine, and histidine, which can produce DBPs [40]. Sunscreens released by swimmers could be the potential precursors of chlorinated and oxidized or nitrogenous DBPs [41].

Country THMs (µg/L) Average or Range HAAs (µg/L) Average or Range References
DBP precursors in swimming pools waters include natural organic matter (NOM) from the filling water, body fluids from the swimmers (e.g., urine, sweat, hair, saliva, etc.), and products used by users (e.g., sunscreen, body lotion, hand soaps, laundry detergents, shampoos, hair gels) [11,29]. In Medellín, the water used to fill pools is from an aqueduct that has total maximum organic carbon values of 1 mg/L. Keuten et al. (2014) found that sweat is the body fluid that provides the second biggest contribution of organic matter to swimming pools [42]. The amount that people sweat in pools depends on both the water temperature and their level of activity. According to Keuten et al. (2014), in cold water, "you don't sweat because the water is cooling your body down and your core temperature doesn't rise" [42], while around 27 or 29 • C, "the cooling effect of the pool water is not enough, and your core temperature starts rising" [42]. In the selected pools of Medellín, the average water temperature was 26.64 • C, the minimum being 23.40 • C and the maximum 32.70 • C. This indicates that pools in Medellin have a significant contribution of sweat by bathers.
The results from pool P9 demonstrate that organic matter is the precursor of THMs. When there are a relatively high number of babies in the pool, the concentration values of THMs increase remarkably compared to when there are few babies in the pool. It is possible that the same occurs with HAAs; however, in the present work this was not very clear.
In some pools (Figure 3), such as P9 and P14, free residual chlorine was close to or slightly greater than 2 mg/L, and higher concentrations of THMs and HAAs were measured than in the other pools. Although in our study the relationship between residual free chlorine and DBPs was only evident in some pools, a direct correlation between free residual chlorine and DBPs has been reported in the literature [26,35,43]. Figure 4 shows that the water temperature in the pools was in the range of 23.40 • C to 32.70 • C during sampling to measure THMs and HAAs, but the pools that exceeded 28 • C, such as P4, P9, and P14, had the highest DBP values. This is in accordance with what was reported by Karanfil et al. (2015) who proved that the increase in residual free chlorine and temperature increases the concentrations of THMs and HAAs [26]. Ilyas et al. (2018) reported that total organic carbon, residual free chlorine, temperature, pH, and bromide play a fundamental role in DBP formation processes [44]. Yang et al. (2016) reported that THM concentrations almost doubled when the temperature increased from 25 to 40 • C [43]. However, the low volatility of HAAs causes their accumulation in the water, causing their concentrations to increase to a very high level over time [45], while the high volatility of THMs prevents their accumulation.
HPs were detected in waters for human consumption in concentrations between 0.010 and 0.486 µg·L −1 in Poland [46], between 0.008 and 0.238 µg·L −1 in Rio de Janeiro [47], and on average 0.06 µg·L −1 in Tunisia [48]. In the present study, we did not detect HPs in swimming pools or water for human consumption. There are very few reports of HPs in swimming pool waters because they are in concentrations that are too low to detect easily. By themselves, HPs are generally in low concentration in water for human consumption. During hypochlorite disinfection, HPs are formed less easily than THMs and HAAs, because hypochlorous acid does not react easily with phenolic compounds, although hypobromous acid does so slightly faster [49]. Despite this, a wide variety of HPs were reported in water for human consumption [50]. Among the few reports on the detection of HPs in swimming pool waters is that of Xiao et al. (2012), who verified the presence of 2,4-dibromophenol, 2,4-dichlorophenol, 2-bromophenol, 2,6-dibromo-4-nitrophenol, 2-bromo-6-chloro-4-nitrophenol, and 2,6-dichloro-4-nitrophenol, which were fully identified using the powerful precursor ion scan method with electrospray ionization triple quadrupole mass spectrometry [51]. They found small amounts of bromide in the chlorine sanitizing agent.

Correlations
In order to determine if there are significant linear relationships between the variables, the correlation matrix was calculated ( Figure 5 and Table 5). It is observed that all relationships are direct, the strongest being between temperature and THMs. HPs were detected in waters for human consumption in concentrations between 0.010 and 0.486 μg·L −1 in Poland [46], between 0.008 and 0.238 μg·L −1 in Rio de Janeiro [47], and on average 0.06 μg·L −1 in Tunisia [48]. In the present study, we did not detect HPs in swimming pools or water for human consumption. There are very few reports of HPs in swimming pool waters because they are in concentrations that are too low to detect easily. By themselves, HPs are generally in low concentration in water for human consumption. During hypochlorite disinfection, HPs are formed less easily than THMs and HAAs, because hypochlorous acid does not react easily with phenolic compounds, although hypobromous acid does so slightly faster [49]. Despite this, a wide variety of HPs were reported in water for human consumption [50]. Among the few reports on the detection of HPs in swimming pool waters is that of Xiao et al. (2012), who verified the presence of 2,4-dibromophenol, 2,4-dichlorophenol, 2-bromophenol, 2,6-dibromo-4-nitrophenol, 2-bromo-6-chloro-4-nitrophenol, and 2,6-dichloro-4-nitrophenol, which were fully identified using the powerful precursor ion scan method with electrospray ionization triple quadrupole mass spectrometry [51]. They found small amounts of bromide in the chlorine sanitizing agent.

Correlations
In order to determine if there are significant linear relationships between the variables, the correlation matrix was calculated ( Figure 5 and Table 5). It is observed that all relationships are direct, the strongest being between temperature and THMs.   Figure 5. Correlation graph. In order to determine the influence of the pool on the concentration of THMs and HAAs, the analysis of variance was performed and results presented in Tables 6 and 7. With a confidence level of 95%, lower p-values (to 0.05) are taken as significant. In conclusion, the HAA concentration values depend on the pool (0.003608 < 0.05), while THM concentration values do not (0.1911 > 0.05). It should be noted that power transformation was performed on the variables to guarantee normality and the assumptions of nonautocorrelation, homoscedasticity, and normality of the residuals were validated.

Conclusions
The pool that has the most THMs and HAAs is P9. This is used for swimming lessons for babies, which indicates that more bodily waste is released into this pool, along with other products, such as sunscreens.
It was observed that the concentration of free residual free chlorine and the temperature favor a higher concentration of THMs and HAAs.
The results are consistent with what was reported previously, that the concentration of HAAs is higher than that of THMs.
The high concentrations of THMs and HAAs found in the public pools of Medellín are certainly not due to the water with which the pools are filled, which is also water for human consumption in Medellín. This is good quality water with very low concentrations of THMs and HAAs, which comply with the maximum permitted values for these compounds.
HPs were not detected, which is consistent with the very few reports confirming the presence of these compounds in swimming pool water.
We recommended that the environmental health authority carry out annual monitoring of DBPs in all the public swimming pools in Medellín, and carry out training with the managers of public swimming pools on how to minimize the presence of DBPs to avoid health risks for pool users.
Author Contributions: Conceptualization, G.A.P. and F.C.; methodology, P.L., V.R., and F.C.; formal analysis, P.L. and V.R.; writing-original draft preparation, P.L.; writing-review and editing, G.A.P. All authors have read and agreed to the published version of the manuscript.