International Journal of

Environmental Research
and Public Health

Article
Evaluation of Physicochemical Characteristics in
Drinking Water Sources Emphasized on Fluoride:
A Case Study of Yancheng, China
Yumin Wang , Ran Yu and Guangcan Zhu *
School of Energy and Environment, Southeast University, Nanjing 210096, China;
wangyumin@seu.edu.cn (Y.W.); yuran@seu.edu.cn (R.Y.)
* Correspondence: gc-zhu@seu.edu.cn; Tel.: +137-7072-6878




Received: 14 December 2018; Accepted: 6 March 2019; Published: 21 March 2019


Abstract: In this study, the concentration of fluoride and the associated health risks for infants,
children, and adults were analyzed and compared for three drinking water sources in Yancheng
City, Jiangsu Province, China. To analyze the relationship between the water quality parameters
of pH, fluoride (F− ), sulfate (SO4 2− ), chloride (Cl− ), total dissolved solids (TDS), total alkalinity
(TAlk), sodium (Na+ ), and potassium (K+ ), statistical analyses including correlation analysis, R-mode
cluster analysis and factor analysis were performed based on monthly data from the year 2010 to
2015. The results indicated: (1) Fluoride concentrations in the drinking water sources ranged from
0.38 to 1.00 mg L−1 (mean = 0.57 mg L−1 ) following the order of Tongyu River > Yanlong Lake >
Mangshe River; (2) fluoride concentrations in 22.93% of the collected samples were lower than
0.5 mg L−1 , which has the risk of tooth cavities, especially for the Mangshe River; (3) the fluoride
exposure levels of infants were higher than children and adults, and 3.2% of the fluoride exposure
levels of infants were higher than the recommended toxicity reference value of 122 µg kg−1 d−1 as
referenced by Health Canada, which might cause dental fluorosis issues; (4) the physico-chemical
characteristics are classified the into four groups reflecting F− - TAlk, Na+ -K+ , SO4 2− -Cl− , and
pH-TDS, respectively, indicating that fluoride solubility in drinking water is TAlk dependent, which
is also verified by R-mode cluster analysis and factor analysis. The results obtained supply useful
information for the health department in Yancheng City, encouraging them to pay more attention to
fluoride concentration and TAlk in drinking water sources.

Keywords: fluoride content; health risk assessment; correlation analysis; cluster analysis; factor
analysis; physicochemical characteristics

1. Introduction
Fluoride (F− ) enters the water body as a result of the geological composition of bedrock and soil,
as well as dental products, food, combustion of fluoride-rich coal, pharmaceuticals, and pesticides. It is
generally accepted that human beings consume fluoride largely from drinking water, and about 90% of
fluoride that exists in drinking water is absorbed by the digestive system [1–4]. Fluoride deficiency or
excess has an adverse effect on human health. Much work had been carried out on the effects of high
fluoride concentration in drinking water, which may cause skeletal issues on both bones and teeth,
especially for children and pregnant women [3–8]. Since fluoride concentration beyond 4 mg L−1
may induce skeletal fluorosis, and beyond 10 mg L−1 may cause crippling skeletal fluorosis, the U.S.
Environmental Protection Agency (EPA) has set the maximum contaminant level (MCL) for fluoride
in drinking water as 4 mg L−1 , and the secondary maximum contaminant level (SMCL) as 2 mg L−1 .
According to Chinese sanitary standards, the fluoride content for drinking water sources is 1.0 mg L−1

Int. J. Environ. Res. Public Health 2019, 16, 1030; doi:10.3390/ijerph16061030 www.mdpi.com/journal/ijerph
Int. J. Environ. Res. Public Health 2019, 16, 1030 2 of 16

(MHPRC, 2006). Although the lower limit of fluoride content in water sources is not specified, fluoride
content of lower than 0.5 mg L−1 will increase the risk of dental caries, affect the formation of dental
enamel, as well as decrease the mineralization of bones, especially for children who are in the body
growth stage [9–12]. Since the 1940s, fluoridation schemes were introduced into the United States,
England, and Wales with the objective of improving dental health, which decreased the proportion of
decayed, missing and filled teeth in children significantly. Additionally, fluoridated toothpaste and
fluoride-containing mouthwash can also contribute to fluoride intake in children [2,13]. The World
Health Organization recommends F− concentrations in drinking water which range from 0.5 mg L−1
to 1.5 mg L−1 for tooth deterioration prevention. The US Public Health Service (USPHS) recently
recommend optimal fluoride concentrations of 0.7 mg L−1 in community water systems, which can
prevent dental caries but also lower the risk of dental fluorosis [12,14]. Since low and high contents of
fluoride in drinking water both pose threats on human health, it is of great importance to evaluate the
F− content in drinking water sources accurately [15,16].
This study aimed to analyze and compare fluoride content and the associated health risks for
infants, children, and adults in three drinking water sources of Yancheng City: the Tongyu River,
Yanlong Lake, and Mangshe River. Moreover, the relationship between the other physicochemical
parameters of pH, fluoride (F− ), sulfate (SO4 2− ), chloride (Cl− ), total dissolved solids (TDS), total
alkalinity (TAlk), sodium (Na+ ), and potassium (K+ ) were investigated, summarized, and analyzed by
multivariate statistical analysis method.

2. Materials and Methods

2.1. Study Area

Yancheng (119◦ 340 –120◦ 270 E, 32◦ 510 –34◦ 120 N) is an eastern coastal city in the center of the Jiangsu
province of China with a population of over 8 million. It faces the Yellow Sea on the east, and is to the
west of Yangzhou City and Huai’an City, the north of Lianyungang City, and the south of Nantong and
Taizhou cities. The research area is 48.54 km2 , shown in Figure 1, with three drinking water sources:
the Tongyu River, Mangshe River, and Yanlong Lake. The Tongyu River has a total length of 415 km,
originating from the Changjiang River, and ends its course in Lianyungang City. The middle reach of
the Tongyu River runs through Yancheng City with a length of 183.6 km, and has an average flow of
approximately 100 m3 /s. The Mangshe River originates from Dazong Lake and discharges into the East
Sea with a total length of nearly 50 km. The Tongyu River and the Mangshe River serve the complex
functions of drinking water sources, fairways, irrigation water sources, and flood drainage system.
The Yanlong Lake is a man-made ecological lake built in 2012 occupying 2.23 km2 with pretreatment
area, ecological wetland purification area, and deep purification area, which also serves as a drinking
water source for Yancheng City [17]. Up to now, several water pollution events have occurred in the
drinking water sources of Yancheng City. In 2009, sewage from chemical plants with volatile phenolic
compounds were illegally drained upstream of water sources, resulting in the cutting off of the water
supply for as long as 66 hours, which severely affected the life of nearly 200,000 inhabitants as well as
industrial processes. Therefore, the water sources of Yancheng city should be monitored closely and
paid more attention to, especially as regards the water quality parameters related to human health.
 Int. J. Environ. Res. Public Health 2019, 16, x 3 of 17
Int. J. Environ. Res. Public Health 2019, 16, 1030 3 of 16

Figure 1. The drinking water sources in Yancheng City.

2.2. Samples and Measurements

One hundred and fifty-six samples, including 60 samples in the Tongyu River from January,
2010 to December, 2015, 54 samples
Figure 1. Theindrinking
the Mangshe River from
water sources July, 2011
in Yancheng City.to Decemeber, 2015, and
42 samples in Yanlong Lake from July, 2012 to December, 2015, were collected during the study.
2.2. GPS
The Samples of sample sites are (120◦ 020 6” E, 33◦ 320 100 N), (120◦ 040 3900 E, 33◦ 210 5600 N), and
and Measurements
coordination
(120◦ 010 3600 E, 33◦ 190 5900 N) for the Tongyu River, Mangshe River, and Yanlong Lake, respectively.
One hundred and fifty-six samples, including 60 samples in the Tongyu River from January,
Parameters of pH, fluoride (F− ), sulfate (SO4 2− ), chloride (Cl− ), total dissolved solids (TDS), total
2010 to December, 2015, 54 samples in the Mangshe River from July, 2011 to Decemeber, 2015, and
alkalinity (TAlk), sodium (Na+ ) and potassium (K+ ) were analyzed using the standard water analysis
42 samples in Yanlong Lake from July, 2012 to December, 2015, were collected during the study. The
methods (seen in Table 1). The selected parameters of pH, SO4 2− , Cl− , TDS, Na+ , and K+ were
GPS coordination of sample sites are (120°02′6′′ E, 33°32′1′′ N), (120°04′39′′ E, 33°21′56′′ N), and
reported to be related to fluoride concentration [5,18,19], and parameters of Ca2+ and Mg2+ had a
(120°01′36′′ E, 33°19′59′′ N) for the Tongyu River, Mangshe River, and Yanlong Lake, respectively.
minor relationship with fluoride- concentration [19]. Moreover, TDS represents the sum of cations and
Parameters of pH, fluoride (F ), sulfate (SO42−), chloride (Cl−), total dissolved solids (TDS), total
anions as the principal inorganic constituents. The parameters were measured in accordance with the
alkalinity (TAlk), sodium (Na+) and potassium (K+) were analyzed using the standard water analysis
standard methods.
methods (seen in Table 1). The selected parameters of pH, SO42−, Cl−, TDS, Na+, and K+ were reported
to be related
Table 1. to fluoride
Limit concentration
specified [5,18,19],
of physicochemical and parameters
parameters of Casources
in drinking water
2+ and Mg2+ had a minor
(CJ3020-93).
relationship with fluoride concentration [19]. Moreover, TDS represents the sum of cations and
Parameters
anions as the principal Limit The
inorganic constituents. Analytical Methods
parameters Detection
were measured Criteria
in accordance with the
standard methods. + (MHPRC, GB/T
Potassium (K ) N/A Flame photometer
5750.6-2006)
Table 1.Sodium
Limit specified
+ of physicochemical parameters in drinking (MHPRC,
water sources GB/T
(CJ3020-93).
(Na ) N/A Flame photometer
5750.6-2006)
Parameters Limit Ion selectivemethods (MHPRC,
Analytical GB/T criteria
Detection
Fluoride (F− ) 1.0
Potassium (K+) N/A electrode
Flame photometer 5750.5-2006)
(MHPRC, GB/T 5750.6-2006)
Sodium (NapH +
) N/A Flame photometer (MHPRC,
(MHPRC, GB/T
GB/T 5750.6-2006)
6.5–8.5 pH-meter
Ion selective 5750.4-2006)
Fluoride (F )
−
1.0 (MHPRC,GB/T
(MHPRC, GB/T 5750.5-2006)
Chloride (Cl− ) 250 electrode
Titrimetric
pH 6.5–8.5 pH-meter 5750.5-2006)
(MHPRC, GB/T 5750.4-2006)
Total Alkalinity
Chloride (Cl−) (TAlk) N/A
250 Titrimetric
Titrimetric (EPBPRC,
(MHPRC,2002)
GB/T 5750.5-2006)
(MHPRC, GB/T 2002)
Total Alkalinity (SO4 2− )
Sulfate(TAlk) N/A
250 Titrimetric
Spectrophotometric (EPBPRC,
5750.5-2006)
Sulfate (SO42−) 250 Spectrophotometric (MHPRC, GB/T 5750.5-2006)
(MHPRC, GB/T
Total Dissolved Solids Solids
Total Dissolved (TDS) (TDS) 1000
1000 Gravimetric
Gravimetric (MHPRC, GB/T 5750.4-2006)
5750.4-2006)
Note 1: Units for all parameters are mg−L1-1 except pH. Note 2: N/A: not available.
Note 1: Units for all parameters are mg L except pH. Note 2: N/A: not available.
Int. J. Environ. Res. Public Health 2019, 16, 1030 4 of 16

2.3. Health Risk Assessment

In order to measure the variation of the health risks caused by exposure to fluoride, the health risk
assessment model was applied to simulate the health effects of fluoride as follows [20]. The population
was divided into three age groups based on physiological and behavioral differences: infants, children,
and adults.
C × IR
ADD = × 1000 (1)
BW
where ADD is exposure to fluoride, expressed as µg kg−1 day−1 , C is the concentration of fluoride in
water (mg L−1 ), IR is water ingestion rate, which is set as 0.75, 1, and 2 L d−1 for infants, children, and
adults, respectively, BW refers to body weight, which is set as 5, 10, and 60 kg for infants, children, and
adults, respectively. A daily intake of 122 µg kg−1 d−1 of fluoride might cause fluorosis, while more
than 200 µg kg−1 d−1 of fluoride might lead to skeletal fluorosis issues [15].
Hazard quotient is assessed by Equation (2) as follows:

ADD
HQ = (2)
RfD

where HQ is hazard quotient of fluoride, RfD is the oral reference dose of fluorine intake in drinking
water, which is adopted to be 60 µg kg−1 d−1 [21]. When the value of HQ is larger than 1.00,
the non-carcinogenic risk will exceed the acceptable level and a hazard is considered to occur.

2.4. Statistical Analysis

To gain insight into the chemical composition of drinking water sources, obtain more information
about dominant indicators influencing the occurrence of fluoride, and analyze the relationships
between fluoride and other physicochemical parameters, the multivariate statistical analysis was
performed with the statistical package for social sciences (SPSS) version 19 (IBM Corporation,
Chicago, USA).
Multivariate statistical analyses including correlation analysis, cluster analysis (CA), principal
component analysis (PCA), and factor analysis (FA) are effective tools for analyzing water quality and
obtaining meaningful results. The CA method is able to group objects into clusters with high internal
(within cluster) homogeneity and high external (between clusters) heterogeneity, usually shown as
a dendrogram (tree diagram), which is a visual display of the analysis results with Dlink/Dmax ×
100 as the horizontal ordinate. Ward’s method with squared Euclidean distance is widely used in CA
to determine the variability. The PCA method extracts principal components by combining original
variables linearly to simplify data sets. FA further simplifies the data structure by rotating the axis
defined by PCA [19]. In this paper, R-mode cluster analysis (R-CA) was applied to distinguish variables
according to the distance between them, and R-mode factor analysis (R-FA) was applied to analyze
and organize the large data sets. The comprehensive scores for drinking water sources were obtained
by Equation (3) as follows:
n p
S = ∑ λi × ( Mr × M ND ) (3)
i =1

where Mr is the rotated factor loadings matrix on all the parameters, which is obtained by the rotated
factor analysis method, M ND is the normalized data matrix, which is obtained from the measured data
in drinking water sources, λi is the eigenvalue of the rotated component matrix, and n is the primary
component number obtained by R-FA.
Int. J. Environ. Res. Public Health 2019, 16, 1030 5 of 16

3. Results and Discussion

3.1. Temporal and Spatial Variations of Fluoride Concentrations

The temporal variations of fluoride concentrations in three drinking water sources are shown
in −1 in the Tongyu
Int. J. Environ.2.Res.
Figure The mean
Public Healthfluoride
2019, 16, x concentrations were 0.58, 0.55, and 0.57 mg L 5 of 17
River, the Mangshe River, and the Yanlong Lake, respectively. In the Tongyu River, the variation
amplitude
Mangshe River, was slight,
and the with monthly
Yanlong mean
Lake, fluoride concentrations
respectively. In the Tongyu ranging
River,from 0.51 to 0.66
the variation mg L−1 .
amplitude
However, − 1
was slight,the minimum
with monthly fluoride concentrations
mean fluoride were lessranging
concentrations than 0.5from mg L0.51intoalmost0.66 mg all L
of−1the months
. However,
except −
0.53 mgfluoride 1
L which occurred inwere October, 2015 0.5andmg 0.51 − 1
the minimum concentrations less than L mg
−1 L which
in almost all ofoccurred
the months in November,
except 0.53
2013.
mg L−1Furthermore,
which occurred the fluoride
in October,concentration
2015 anddropped0.51 mgtoL−1 aswhich 0.38 mg Lin−1November,
low asoccurred in January, 2013.
2010.
The low concentration of fluoride may induce dental caries. Similarly,
Furthermore, the fluoride concentration dropped to as low as 0.38 mg L in January, 2010. The low −1 the maximum and mean
fluoride
concentration concentrations
of fluoride alsomayreached
induce higher
dental values
caries.in Similarly,
October and the November,
maximum 2010, and meanwhichfluoride
means
that fluoride contents
concentrations may behigher
also reached associated
values with groundwater
in October with high 2010,
and November, fluoride concentration
which means thatsince the
fluoride
river is supplied by groundwater in dry seasons (October and November).
contents may be associated with groundwater with high fluoride concentration since the river is The fluoride of shallow
groundwater
supplied by in Yancheng City
groundwater in was
dry reported
seasons to(October
exceed the andstandard
November). level ofThe1.0 mg L−1 , which
fluoride led to
of shallow
dental and bone fluorosis [22]. Additionally, a rising number of fluorine-related
groundwater in Yancheng City was reported to exceed the standard level of 1.0 mg L , which led chemical−1 plants wereto
constructed
dental and bone in the Yancheng
fluorosis [22].districts, whicha may
Additionally, risingincrease
numberthe fluoride contentchemical
of fluorine-related and corresponding
plants were
health
constructedrisks through irregular discharge
in the Yancheng districts, of wastewater
which into surrounding
may increase the fluoride drinking
contentwater sources. As for
and corresponding
the Mangshe River, the variety range of maximal, minimal, and mean
health risks through irregular discharge of wastewater into surrounding drinking water sources. fluoride content remained almostAs
invariable,
for the Mangshe whichRiver,
suggeststhe that fluoride
variety rangein ofthe Mangshe
maximal, River may
minimal, come from
and mean steady
fluoride contentgroundwater.
remained
However, it is worth
almost invariable, noting
which that thethat
suggests maximal
fluoride fluorides
in the were
Mangshe lowest (<0.5may
River L−1 ) in
mg come May,steady
from 2012
and June, 2015, and minimal fluorides were also lowest (<0.5 mg L −1 ) in May, 2015 and June, 2013.
groundwater. However, it is worth noting that the maximal fluorides were lowest (<0.5 mg L ) in −1

The
May,mean 2012fluoride
and June, content
2015,reached its maximum
and minimal fluorides value
wereofalso mg L−1(<0.5
0.64 lowest in August.
mg L−1)This supported
in May, 2015 andthe
claim that large quantities of upstream agricultural nonpoint sources
June, 2013. The mean fluoride content reached its maximum value of 0.64 mg L in August. This and domestic pollution
-1 sources
were
supporteddrained theinto the that
claim Mangshe
largeRiver, especially
quantities in flood agricultural
of upstream seasons. Thenonpoint
differences between
sources and thedomestic
Tongyu
River
pollution andsources
the Mangshe River could
were drained be due
into the Mangsheto theRiver,
fluoride in the Mangshe
especially River being
in flood seasons. The affected
differences by
anthropogenic activity more seriously than the Tongyu River. As for
between the Tongyu River and the Mangshe River could be due to the fluoride in the Mangshe River the Yanlong Lake, the maximum
fluoride
being affectedcontent by in July, 2012 and August,
anthropogenic activity 2014 reached
more 0.80 and
seriously than0.95the L−1 , respectively.
mg Tongyu River. As Since
for July
the
and
YanlongAugust arethe
Lake, in flood seasons,
maximum large quantities
fluoride con of wastewater probably gave rise to fluoride in the
Yanlong Lake.

Mean
Fluoride concentration in Tongyu River (mg L )
-1

0.7 Standard deviation

0.6

0.5

0.4

0.3

0.2

0.1

0.0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
(a)

Figure 2. Cont.
Int. J. Environ. Res. Public Health 2019, 16, 1030 6 of 16
Int. J. Environ. Res. Public Health 2019, 16, x 6 of 17

Mean
Standard deviation

Fluoride concentration in Mangshe River (mg L )

0.7

-1
0.6

0.5

0.4

0.3

0.2

0.1

0.0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month
(b)
Mean
Fluoride concentration in Yanlong Lake (mg L )
-1

0.8 Standard deviation

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month

(c)
Figure2.2.Temporal
Figure Temporalvariations
variationsofoffluoride
fluorideconcentrations
concentrationsininthree
threedrinking
drinkingwater
watersources
sources(a)
(a)Monthly
Monthly
fluoride concentration in the Tongyu River. (b) Monthly fluoride concentration in the Mangshe
fluoride concentration in the Tongyu River. (b) Monthly fluoride concentration in the Mangshe River. River.
(c) Monthly fluoride concentration in the Yanlong
(c) Monthly fluoride concentration in the Yanlong Lake. Lake.

Thedistributions
The distributionsofoffluoride
fluorideconcentration
concentrationin indrinking
drinkingwater
watersourcessourcesare
arepresented
presentedin inFigure
Figure3.3.
InIn157
157representative
representativewater
watersamples,
samples,including
including68 68samples
samplesin inthe
theTongyu
TongyuRiver,
River,49 49samples
samplesininthethe
Mangshe River, and 40 samples in the Yanlong Lake (Figure 3a), the fluoride
Mangshe River, and 40 samples in the Yanlong Lake (Figure 3a), the fluoride contents were less than contents were less than
theChinese
the Chinesesanitary
sanitarystandard
standardfor fordrinking
drinkingwaterwater(1.0(1.0mg mgL−L1−1),), the
theWHO
WHOdrinking
drinkingwaterwaterquality
quality
guideline(1.50
guideline (1.50 mg − 1
mg L−1) and
) andU.S. standard
U.S. standard (2.0(2.0
mgmg − 1
L−1).LHowever,
). However,in moderate samples
in moderate (22.9%,(22.9%,
samples n = 157)
nincluding 12 samples
= 157) including in the Tongyu
12 samples River, 16River,
in the Tongyu in the16Mangshe River, and
in the Mangshe 8 in and
River, the Yanlong Lake, the
8 in the Yanlong
fluoride
Lake, the concentrations were less were
fluoride concentrations than 0.5
lessmg
thanL−10.5 mg L−1 recommended
recommended by the WHO byto prevent
the WHO tooth cavity
to prevent
(seen in Figure 3b). This may pose an elevated risk of dental decay, especially for
tooth cavity (seen in Figure 3b). This may pose an elevated risk of dental decay, especially for children.children. Moreover,
the data for
Moreover, thefluoride concentration
data for fluoride refer to refer
concentration drinking water sources,
to drinking water sources,not treated finished
not treated waterwater
finished from
drinking
from water
drinking treatment
water treatmentplants, where
plants, fluoride
where fluoride maymaybe beremoved
removed bybycoagulation–sedimentation
coagulation–sedimentationin
inthe
thetraditional
traditionalpurification
purificationprocess.
process.The Theremoval
removalefficiency
efficiency depends
depends on on the type
type ofof coagulants,
coagulants,
dosage, and sedimentation time. The removal of fluoride increased the risk of dental decay for
children. For preventing dental decay, adding fluoride to drinking water is an effective measure. In
Int. J. Environ. Res. Public Health 2019, 16, 1030 7 of 16

dosage, and sedimentation time. The removal of fluoride increased the risk of dental decay for
Int. J. Environ. Res. Public Health 2019, 16, x 7 of 17
children. For preventing dental decay, adding fluoride to drinking water is an effective measure. In the
U.S.,the
theU.S.,
prevalence rate of dental
the prevalence rate ofcaries
dentalhas dropped
caries by 50%–60%
has dropped in the last
by 50%–60% 30 years
in the last 30asyears
a result
as aof
result
fluoridation in drinking water. Similar operations could be found in other countries
of fluoridation in drinking water. Similar operations could be found in other countries such assuch as England,
Russia, Brazil, Russia,
England, and Cuba. Therefore,
Brazil, local Therefore,
and Cuba. governments should
local take measures
governments shouldto take
prevent dental to
measures decay,
prevent
suchdental
as promoting the use of toothpaste with fluoride.
decay, such as promoting the use of toothpaste with fluoride.

(a)

(b)
FigureFigure 3. Fluoride
3. Fluoride contents
contents and and frequency
frequency distribution
distribution in drinking
in drinking water
water sources
sources of of Yancheng
Yancheng City (a)
City
Fluoride
(a) Fluoride contents
contents in in drinking
drinking water
water sources
sources of Yancheng
of Yancheng City.City.
Note:Note: Dashed
Dashed line refers
line refers to fluoride
to fluoride
concentration
concentration of 0.5ofmg −1 ,Lwhich
0.5Lmg −1, which is the limit to prevent tooth cavity recommended by the WHO. (b)
is the limit to prevent tooth cavity recommended by the WHO.
Frequency
(b) Frequency distributions
distributions ofof fluoride
fluoride concentration
concentration inin the
the studied
studied area.
area.

3.2. Health Risk Assessment of Fluoride

The exposure levels of fluoride (ADD) were calculated using Equation (1), and shown in Table
S1 (Supporting information). The fluoride exposure level for infants in all the samples were analyzed
and shown in Figure 4. The infants of 4 of 68 (5.9%), 0 of 49 (0%), and 1 of 40 (2.5%) for the Tongyu
River, Mangshe River, and Yanlong Lake, respectively, had fluoride exposure levels higher than the
Int. J. Environ. Res. Public Health 2019, 16, 1030 8 of 16

3.2. Health Risk Assessment of Fluoride

The exposure levels of fluoride (ADD) were calculated using Equation (1), and shown in Table S1
(Supporting information). The fluoride exposure level for infants in all the samples were analyzed and
shownInt.in FigureRes.
J. Environ. 4. Public
The infants of 16,
Health 2019, 4 of
x 68 (5.9%), 0 of 49 (0%), and 1 of 40 (2.5%) for the Tongyu 8River,of 17
Mangshe River, and Yanlong Lake, respectively, had fluoride exposure levels higher than the toxicity
reference
toxicityvalue (TRVs)value
reference as recommended by Health Canada
(TRVs) as recommended (122
by Health Canada −1 dμg
µg kg(122 −1 ),kg
which
−1 d−1),can cause
which candental
cause
dental
fluorosis fluorosis issues.
issues.

Figure 4. Exposure of fluoride for infants in the three drinking water sources of Yancheng City. Note:
Figure 4. Exposure of fluoride for infants in the three drinking water sources of Yancheng City. Note:
Dashed line in Figure 4 refers to the toxicity reference value of 122 µg kg−1 d−1 , which is recommended
Int. J. Environ. Res. Public
Dashed line inHealth
Figure2019, 16, x to the toxicity reference value of 122 μg kg−1 d−1, which is recommended
4 refers 9 of 17
by Health Canada.
by Health Canada.

The hazard quotients were calculated using Equation (2) and shown in Figure 5. All the hazard
quotients of infants were higher than the recommended value of 1.0 by United States Environmental
Protection Agency (USEPA). The results suggested that the infants in Yancheng City suffer from the
health risks associated with fluoride. Compared with the Tongyu River and Mangshe River, the
hazard quotients were the biggest in Yanlong Lake in 2012 and 2013. In 2012, the hazard quotients
showed a significant difference among the three drinking water sources. However, in 2015, the
hazard quotients of the three drinking water sources were similar. In addition, the hazard quotients
of children and adults were almost always lower than 1.0 except in Yanlong Lake in 2012, which
means that the children and adults would not suffer from the harmful effects associated with daily
fluoride intake.

Figure5.5.Hazard
Figure Hazard quotient
quotient of fluoride
fluoride in
in the
thethree
threedrinking
drinkingwater
water sources
sources from 2012
from to 2015.
2012 Note:
to 2015. Note:
Dashedline
Dashed lineininFigure
Figure 55 refers
refers to the hazard
hazardquotient
quotient(HQ)
(HQ)ofof1.0,
1.0,which
whichis is
thethe
limit recommended
limit recommended by by
theWHO.
the WHO.

3.3. Statistical Analysis

3.3.1. Concentrations of Other Physic-Chemical Parameters

The concentration ranges of the other physicochemical parameters are shown in Table 2. The pH
Int. J. Environ. Res. Public Health 2019, 16, 1030 9 of 16

The hazard quotients were calculated using Equation (2) and shown in Figure 5. All the hazard
quotients of infants were higher than the recommended value of 1.0 by United States Environmental
Protection Agency (USEPA). The results suggested that the infants in Yancheng City suffer from the
health risks associated with fluoride. Compared with the Tongyu River and Mangshe River, the hazard
quotients were the biggest in Yanlong Lake in 2012 and 2013. In 2012, the hazard quotients showed a
significant difference among the three drinking water sources. However, in 2015, the hazard quotients
of the three drinking water sources were similar. In addition, the hazard quotients of children and
adults were almost always lower than 1.0 except in Yanlong Lake in 2012, which means that the
children and adults would not suffer from the harmful effects associated with daily fluoride intake.

3.3. Statistical Analysis

3.3.1. Concentrations of Other Physic-Chemical Parameters

The concentration ranges of the other physicochemical parameters are shown in Table 2. The pH
values ranged from 7.2 to 8.4 with an average of 7.6 in the Tongyu River, from 7.3 to 8.7 with an
average of 7.7 in the Mangshe River, and from 7.4 to 8.5 with an average of 8.1 in Yanlong Lake.
Compared with the WHO recommended limits of pH 6.5–8.5, the drinking water sources of Yancheng
City are alkaline. The TDS levels of the three drinking water sources ranged from 122 to 594 mg L−1 ,
less than the maximum contamination limit (MCL) of 2000 mg L−1 recommended by the WHO
(shown in Table 1). However, some of them exceeded the TDS levels of 500 mg L−1 , which may be
unacceptable for consumers because of its taste. High intake of sodium (Na+ ) is harmful for patients
suffering from renal, cardiac, and circulatory disease. In Table 2, the concentrations of Na+ are less than
200 mg L−1 , and amounts of K+ are less than 250 mg L−1 , which is listed as potential complaint for
taste. The concentrations of Cl− ranged from 35 to 194 mg L−1 , less than the limit value of 250 mg L−1
to prevent heart and kidney diseases from a health perspective. The SO4 2− concentration in drinking
water sources ranged from 23 to 81 mg L−1 , and high concentration of SO4 2− may contribute to the
taste and the corrosion of water supply distribution systems [5]. The relationship between pH and
dissolution of fluoride can be obtained from the Mangshe River and Yanlong Lake. In the Mangshe
River, the highest pH value of 8.72 occurred in February, 2012, while the concentration of fluoride was
also at high level of 0.63 mg L−1 . The lowest pH value of 7.28 occurred in September, 2015, while the
concentration of fluoride was also at lower level of 0.50 mg L−1 . Similarly, in Yanlong Lake, the highest
pH value of 8.54 occurred in August, 2012, while the concentration of fluoride reached 0.73 mg L−1 .
The lowest pH value of 7.42 occurred in November, 2013, while the concentration of fluoride was also
at lower level of 0.49 mg L−1 . However, in the Tongyu River, the highest pH value of 8.40 occurred in
February, 2012, while the concentration of fluoride was only 0.53 mg L−1 . The lowest pH value of 7.20
occurred in February, 2013, while the concentration of fluoride was also at lower level of 0.49 mg L−1 .
Generally, the results comply with the conclusion that F− contents are associated with the pH values,
which is obtained from other literature [23]. The concentrations of Na+ in the Tongyu River are higher
than that in the Mangshe River and Yanlong Lake, which is the same result obtained from fluoride.
The results indicate that higher Na+ concentration increases the dissolution of fluoride, which is similar
to the findings from other literature [18].
Int. J. Environ. Res. Public Health 2019, 16, 1030 10 of 16

Table 2. Other physicochemical parameters in three drinking water sources of Yancheng City.
Int. J. Environ. Res. Public Health 2019, 16, x 10 of 17
Drinking Water Sources pH SO4 2− Cl− TDS TAlk Na+ K+
Max. 8.40 81 194
the pH values, whichTongyu
is obtained
River
from other7.20
Min.
literature
23
[23].
59
The594 218 85
concentrations
269 117 44
12
of1.2Na+ in the Tongyu
River are higher than that in the Mangshe
Mean River
7.57 and49Yanlong
114 Lake,
411 which
167 is 65
the same
6.9 result obtained
from fluoride. The results indicate thatMax.
higher Na+ concentration
8.72 64 116 increases
523 187 the90 dissolution
7.4 of fluoride,
Mangshe River Min. 7.28 25 35 122 123 34 1.6
which is similar to the findings from other
Mean literature
7.65 43 [18]. 69 380 158 60 5.7
Max. 8.54 61 97 512 183 89 9
Table 2. OtherYanlong
physicochemical
Lake parameters
Min. 7.42 in three
24 drinking
46 water sources
231 116 of Yancheng
34 1.6 City.
Mean 8.12 44 71 369 156 60 6.4
Drinking water sources pH are
Note: Units for all parameters SO Cl pH.TDS
4 L−1 except
mg
2− − TAlk Na+ K+
Max. 8.40 81 194 594 218 85 12
3.3.2. Correlation Analysis
Tongyu River Min. 7.20 23 59 269 117 44 1.2
Pearson correlation coefficients Meanfor the possible 7.57 49 114 between
relationships 411 167
fluoride 65 the other
and 6.9
Max. 8.72 64 116 523 187 90 7.4
monitored physicochemical parameters are summarized in Table 3. There is a positive relationship
Mangshe River Min. 7.28 25 35 122
between fluoride and total alkalinity (TAlk) (r = 0.53), suggesting that fluoride content is associated 123 34 1.6
with the increase of alkalinity inMean surface water,7.65 which is43similar69to results 380 from158 60
other literature 5.7
[24].
Max. 8.54 61 97 512 183 89 9
The concentrations of TAlk in the Tongyu River are higher than the Mangshe River and Yanlong Lake,
Yanlong Lake Min. 7.42 24 46 231 116 34 1.6
where the concentration of fluoride is also the highest. The correlation between fluoride and TAlk is
Mean 8.12 44 71 369 156 60 6.4
shown in Figure 6a. Besides TAlk, the statistical relationship between fluoride concentration and Cl− is
Note: Units for all parameters are mg L−1 except pH.
shown in Figure 6b with positive correlation coefficients of 0.42. The chlorine contents in three drinking
water sources are shown in Figure 6c. Similar to fluoride, the chlorine contents in the Tongyu River are
3.3.2. Correlation Analysis
obviously higher than those in the Mangshe River and Yanlong Lake. As for the other parameters,
thePearson correlation coefficients forpH,
theSO 2−
possible
relationship between fluoride and 4 , andrelationships
TDS are positive betweenwith fluoride and
correlation the other
coefficients
monitored
of 0.39, 0.22,physicochemical
and 0.11, respectively.parameters are summarized
Moreover, the relationship in Table
between3. There is a positive
fluoride + and K+ are
and Na relationship
between with
negative fluoride and total
correlation alkalinityof
coefficients (TAlk)
−0.15(rand = 0.53),
−0.47, suggesting that fluoride content is associated
respectively.
with the increase of alkalinity in surface water, which is similar to results from other literature [24].
The concentrations ofTable TAlk3.inCorrelation
the Tongyu matrix
Riverof parameters
are higherinthan
the studied area.
the Mangshe River and Yanlong
Lake, where the concentration
Parameters pH
of fluoride
F−
is
SO4 2−
also the
Cl−
highest.
TDS
The correlation
TAlk Na+
between
K+
fluoride and
TAlk is shown in Figure pH 6a. 1.00
Besides TAlk, the statistical relationship between fluoride concentration
F−
and Cl- is shown in Figure 6b
0.39with 1.00
positive correlation coefficients of 0.42. The chlorine contents in
SO 2− 0.19 0.22 1.00
three drinking waterCl4sources
− are
−0.23 shown
0.42 in0.52
Figure
* 6c. Similar to fluoride, the chlorine contents in the
1.00
Tongyu River are obviously higher than those in the
TDS 0.09 0.11 0.22 − 0.06 Mangshe
1.00 River and Yanlong Lake. As for the
TAlk −0.17 0.53 * −0.12 0.44 −0.22 1.00
other parameters, the Na + relationship
−0.46 between
−0.15 −0.05 fluoride
0.24 and
−0.31pH, SO42−, and
0.32 1.00 TDS are positive with
+
correlation coefficients K of −0.39, −0.47 and−0.18
0.32 0.22, 0.00
0.11, respectively. −0.07 Moreover,
−0.10 0.68
the 1.00
* relationship between
Note: “*” refers to relative coefficients higher than 0.5.
fluoride and Na and K are negative with correlation coefficients of −0.15 and −0.47, respectively.
+ +

(a)

Figure 6. Cont.
Int. J. Environ. Res. Public Health 2019, 16, 1030 11 of 16
Int. J. Environ. Res. Public Health 2019, 16, x 11 of 17

(b)

(c)
Figure
Figure 6. Relationships
6. Relationships between
between F- and
F− and TAlk,
TAlk, F−F and
- and chloride, and chlorine contents in drinking water
chloride, and chlorine contents in drinking
watersources
sourcesofofYancheng
YanchengCity
City(a)
(a)Correlation betweenFFand
Correlationbetween TAlk (b)(b)
Correlation between F- and chloride
- −
and TAlk Correlation between F− and
(c) Chlorine
chloride contents
(c) Chlorine in drinking
contents water
in drinking sources
water of Yancheng
sources City.
of Yancheng City.

Besides the relationship between fluoride and alkalinity, the correlation analyses of other
parameters are described as follows: (1) Chloride is related to sulfate with a correlation coefficient
of 0.52; (2) sodium (Na+ ) is observed to have a close relationship with potassium (K+ ) (r = 0.68).
The correlation was further interpreted through R-mode cluster analysis and factor analysis.

3.3.3. R-mode Cluster Analysis and Factor Analysis

The R-mode cluster analysis was carried out for different physicochemical characters, and the
results are shown in a Figure 7. All the variables are grouped into four clusters in a convincing way
of Dlink/Dmax × 100 < 10. Fluoride (F− ) and total alkalinity (TAlk) were grouped into a cluster.
The correlation between F− and TAlk is also shown in Table 3 and Figure 6a with a correlation
coefficient of 0.53. In addition, the correlation between sodium (Na+ ) and potassium (K+ ) is shown in
Table 3 and Figure 8a with a correlation coefficient of 0.68. Similarly, the correlation between sulfate
(SO4 2− ) and chloride (Cl− ) is shown in Table 3 and Figure 8b with a correlation coefficient of 0.52.
of Dlink/Dmax × 100 < 10. Fluoride (F-) and total alkalinity (TAlk) were grouped into a cluster. The
correlation between F- and TAlk is also shown in Table 3 and Figure 6a with a correlation coefficient
of 0.53. In addition, the correlation between sodium (Na+) and potassium (K+) is shown in Table 3 and
Figure 8a with a correlation coefficient of 0.68. Similarly, the correlation between sulfate (SO42−) and
Int. J. Environ. Res.
chloride (Cl-) isPublic Health 2019, 16, 1030
shown in Table 3 and Figure 8b with a correlation coefficient of 0.52. 12 of 16

Int. J. Environ. Res. Public Health 2019, 16, x 13 of 17

Figure 7. R-Cluster analysis results.
Figure 7. R-Cluster analysis results.

(a)

(b)

Figure Relationships between

Figure 8. Relationships between Na Na++ and K++, ,and SO442−2−
andSO and
and ClCl − (a) Correlation between Na
− (a) + and+ K+ .
Correlation between Na+ and K . (b)
(b) Correlation between 2 − −
Correlation between SO4SOand
2−
4 and. Cl .
Cl −

The qualitative information about clustering behaviors was extracted by R-mode factor analysis.
The eigenvalues and cumulative variability of all the eight factors are plotted in Figure 9. The first
four factors, having a total variance of 83.31% from all the parameters, correspond to four clusters
obtained from the cluster analysis.
Int. J. Environ. Res. Public Health 2019, 16, 1030 13 of 16

The qualitative information about clustering behaviors was extracted by R-mode factor analysis.
The eigenvalues and cumulative variability of all the eight factors are plotted in Figure 9. The first four
factors, havingRes.
Int. J. Environ. a total
Publicvariance of16,
Health 2019, 83.31%
x from all the parameters, correspond to four clusters obtained
14 of 17
from the cluster analysis.

Figure 9. Scree plot of varimax rotated factors.

Figure 9. Scree plot of varimax rotated factors.
In addition, the rotated factor loadings matrix on all the parameters is expressed in Table 4,
including eigenvalues,
In addition, the percentage of variance,
rotated factor loadingsand cumulative
matrix percentage.
on all the parameters is expressed in Table 4,
including eigenvalues, percentage of variance, and cumulative percentage.
Table 4. Varimax rotated R-mode factor loadings on all the parameters.
Table 4. Varimax rotatedVF1
Parameters R-mode factor
VF2 loadings
VF3on all the VF4
parameters

Parameters
pH −0.80 * VF1 −0.14 VF2
0.15 −VF3
0.19 VF4
F − − 0.45 0.75 * 0.25 0.04
pH −0.80 * −0.14 0.15 −0.19
2− − 0.15 − 0.06 0.13
SO
F4- −0.45 0.75 * *
0.94 0.25 0.04
Cl − 0.24 0.55 0.70 * − 0.01
SO4 2- −0.15 −0.06 0.94 * 0.13
TDS
Cl - − 0.09 −
0.24 0.09 0.12
0.55 0.70 **
0.94 −0.01
TAlk 0.16 0.91 * −0.07 −0.16
TDS + −0.09 −0.09 0.12 0.94 *
Na 0.81 * 0.14 0.12 −0.35
TAlk+ 0.16 0.91 * −0.07 −0.16
K 0.78 * −0.32 0.01 −0.15
Na+
Eigenvalue 2.22 0.81 1.85* 0.14
1.49 0.12
1.10 −0.35
% ofKthe total 0.78 * −0.32 0.01 −0.15
+

Eigenvalue 27.77 2.2223.17 18.60

1.85 13.77
1.49 1.10
variance
% of%
the
oftotal variance
cumulative 27.77 27.77 50.94 23.17
69.54 18.60
83.31 13.77
% of cumulative 27.77
Note: “*” refers to absolute 50.94
values higher than 0.7. 69.54 83.31
Note: “*” refers to absolute values higher than 0.7.
In Table 4, Factor 1 accounted for 27.77% of the total variance and had high loadings of pH (−0.80),
+ In Table 4,+Factor 1 accounted for 27.77% of the total variance and had high loadings of pH
Na (0.81) and K (0.78), +which indicate that the presence of Na+ and K + decreased the values of
(−0.80), Na (0.81) and K (0.78), which indicate that the presence of Na and K+ decreased the values
+ +
pH. Factor 2, accounting for 23.17% of the total variance, was mainly associated with significantly
of pH. Factor 2, accounting for 23.17% of the total variance, was mainly associated with−significantly
higher loading of TAlk (0.91) and moderate loading of F− (0.75), which illustrated that F solubility
higher loading of TAlk (0.91) and moderate loading of F- (0.75), which illustrated that F- solubility
− (0.70),is
is TAlk dependent. Factor 3 was an index of the main anions, including SO42−2− (0.94) and Cl
TAlk dependent. Factor 3 was an index of the main anions, including SO4 (0.94) and Cl (0.70), and
−
and counted for 18.6% of the total variability, which indicated that the SO2−4 2− values of those samples
counted for 18.6% of the total variability, which indicated that the SO4 values of those samples are
are higher when there is an elevated amount of Cl−− . Factor 4, counting for 13.77% of the total variability,
higher when there is an elevated amount of Cl . Factor 4, counting for 13.77% of the total variability,
had a strong relationship with TDS (0.94), which reflected the amount of cations and anions.
had a strong relationship with TDS (0.94), which reflected the amount of cations and anions.
The comprehensive scores of water sources were calculated using Equation (3), and shown in
The comprehensive scores of water sources were calculated using Equation (3), and shown in
Figure 10. It is evident that the comprehensive scores of the Tongyu River are significantly greater
Figure 10. It is evident that the comprehensive scores of the Tongyu River are significantly greater
than those of the Mangshe River and Yanlong Lake, which is due to the fact that the hydrogeological
than those of the Mangshe River and Yanlong Lake, which is due to the fact that the hydrogeological
conditions in the Tongyu River were quite different from that in the Mangshe River and Yanlong Lake.
conditions in the Tongyu River were quite different from that in the Mangshe River and Yanlong
Lake.
Int. J. Environ. Res. Public Health 2019, 16, 1030 14 of 16
Int. J. Environ. Res. Public Health 2019, 16, x 15 of 17

Figure 10.
Figure Comprehensive score
10. Comprehensive score of
of drinking
drinking water
water sources.
sources.

4. Conclusions
4. Conclusions
This study provided a systematic analysis of fluoride contents in the three drinking water sources
This study provided a systematic analysis of fluoride contents in the three drinking water
of Yancheng City in the Jiangsu province of China. The fluoride concentrations in the drinking water
sources of Yancheng City in the Jiangsu province of China. The fluoride concentrations in the
sources were between 0.38 and 1.0 mg L−1 , which were lower than the recommended value in either
drinking water sources were between 0.38 and 1.0 mg L−1, which were lower than the recommended
the local or international drinking water sources quality guidelines, which means that local people
value in either the local or international drinking water sources quality guidelines, which means that
would not suffer from dental fluorosis. The mean fluoride concentrations decreased in the order
local people would not suffer from dental fluorosis. The mean fluoride concentrations decreased in
of Tongyu River > Yanlong Lake > Mangshe River. The results also indicated that the health risks
the order of Tongyu River > Yanlong Lake > Mangshe River. The results also indicated that the health
associated with fluoride in infants are higher than children and adults, which might cause dental
risks associated with fluoride in infants are higher than children and adults, which might cause dental
fluorosis issues, especially for the Tongyu River and Yanlong Lake. However, the fluoride contents
fluorosis issues, especially for the Tongyu River and Yanlong Lake. However, the fluoride contents
less than 0.50 mg L−1 unexpectedly occurred in 22.93% of the samples, including 12 samples in
less than 0.50 mg L−1 unexpectedly occurred in 22.93% of the samples, including 12 samples in the
the Tongyu River, 16 in the Mangshe River and 8 in Yanlong Lake, which increase the risk of tooth
Tongyu River, 16 in the Mangshe River and 8 in Yanlong Lake, which increase the risk of tooth
cavities. Therefore, continuous and closer monitoring work should be performed by the local water
cavities. Therefore, continuous and closer monitoring work should be performed by the local water
management department. In addition, to prevent dental caries, corresponding measures should
management department. In addition, to prevent dental caries, corresponding measures should be
be taken to increase fluoride content through tea, fluoridated toothpaste, and fluoridated salt, etc.,
taken to increase fluoride content through tea, fluoridated toothpaste, and fluoridated salt, etc.,
especially for the Mangshe River.
especially for the Mangshe River.
The correlation analysis among other physicochemical parameters indicated that relationships
The correlation analysis among other physicochemical parameters indicated that relationships
existed between F−− and TAlk (r = 0.53), 2− SO4 2− and Cl− (r = 0.52), and Na+ and K+ (r = 0.68),
existed between F and TAlk (r = 0.53), SO4 and Cl− (r = 0.52), and Na+ and K+ (r = 0.68), which was
which was also verified by R-mode cluster analysis and factor analysis. Four factors were extracted
also verified by R-mode cluster analysis and factor analysis. Four factors were extracted by R-mode
by R-mode factor analysis, which represent cation, fluoride content, anion, and TDS, respectively.
factor analysis, which represent cation, fluoride content, anion, and TDS, respectively. The
The concentrations of TAlk in the Tongyu River were higher than Yanlong Lake and the Mangshe
concentrations of TAlk in the Tongyu River were higher than Yanlong Lake and the Mangshe River,
River, which is in accordance with the fluoride level in three drinking water sources. The results
which is in accordance with the fluoride level in three drinking water sources. The results of
of comprehensive assessment illustrated that physicochemical character of the Tongyu River was
comprehensive assessment illustrated that physicochemical character of the Tongyu River was
significantly different from that of the Mangshe River and Yanlong Lake, which is primarily due to
significantly different from that of the Mangshe River and Yanlong Lake, which is primarily due to
different hydrogeological conditions.
different hydrogeological conditions.
Supplementary Materials: The following are available online at http://www.mdpi.com/1660-4601/16/6/1030/
Author
s1. Contributions: Writing-Review & Editing, Y.W.; Investigation, R.Y.; Supervision, G.Z.

This work wasWriting-Review

Author Contributions:
Funding: & Editing,
funded by the Special S&TY.W.; Investigation,
Project R.Y.;and
on Treatment Supervision,
Control ofG.Z.
Water Pollution from
Bureau of Housing and Urban-rural Development of Jiangsu Province numbered as 2014ZX07405002.
Funding: This work was funded by the Special S&T Project on Treatment and Control of Water Pollution from
Bureau ofof
Conflicts Housing and
Interest: Urban-rural
The Development
authors declare of of
no conflict Jiangsu Province numbered as 2014ZX07405002.
interest.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Abouleish, M.Y. Evaluation of fluoride levels in bottled water and their contribution to health and teeth
problems in the United Arab Emirates. Saudi Dent. J. 2016, 28, 194–202.
Int. J. Environ. Res. Public Health 2019, 16, 1030 15 of 16

References
1. Abouleish, M.Y. Evaluation of fluoride levels in bottled water and their contribution to health and teeth
problems in the United Arab Emirates. Saudi Dent. J. 2016, 28, 194–202. [CrossRef]
2. Chuah, C.J.; Lye, H.R.; Ziegler, A.D.; Wood, S.H.; Kongpun, C.; Rajchagool, S. Fluoride: A naturally-occurring
health hazard in drinking-water resources of Northern Thailand. Sci. Total Environ. 2016, 545–546, 266–279.
[CrossRef]
3. Gautam, R.; Bhardwaj, N.; Saini, Y. Dental fluorosis—A case study from Nawa tehsil in Nagaur district,
Rajasthan (India). Environmentalist 2011, 31, 401–406. [CrossRef]
4. Yousefi, M.; Ghoochani, M.; Mahvi, A.H. Health risk assessment to fluoride in drinking water of rural
residents living in the Poldasht city, Northwest of Iran. Ecotoxicol. Environ. Saf. 2018, 148, 426–430. [CrossRef]
[PubMed]
5. Chandio, T.A.; Khan, M.N.; Sarwar, A. Fluoride estimation and its correlation with other physicochemical
parameters in drinking water of some areas of Balochistan, Pakistan. Environ. Monit. Assess. 2015, 187,
531–539. [CrossRef]
6. Fang, J.; Wu, X.; Xu, J.; Yang, X.; Song, X.; Wang, G.; Yan, M.; Yan, M.; Wang, D. Water management challenges
in the context of agricultural intensification and endemic fluorosis: The case of Yuanmou County. Ecohealth
2011, 8, 444–455. [CrossRef]
7. Jianmin, B.; Yu, W.; Juan, Z. Arsenic and fluorine in groundwater in western Jilin Province, China: Occurrence
and health risk assessment. Nat. Hazards 2015, 77, 1903–1914. [CrossRef]
8. McGrady, M.G.; Ellwood, R.P. Patcharawan Srisilapanan, Narumanas Korwanich, Helen V. Worthington &
Lain Pretty. Dental fluorosis in populations from Chiang Mai, Thailand with different fluoride exposures—
Paper 1: Assessing fluorosis risk, predictors of fluorosis and the potential role of food preparation. BMC
Oral Health 2012, 12, 16–27. [PubMed]
9. Dehbandi, R.; Moore, F.; Keshavarzi, B. Geochemical sources, hydrogeochemical behavior, and health risk
assessment of fluoride in an endemic fluorosis area, central Iran. Chemosphere 2018, 193, 763–776. [CrossRef]
[PubMed]
10. Ozsvath, D.L. Fluoride and environmental health: A review. Rev. Environ. Sci. Biotechnol. 2008, 8, 59–79.
[CrossRef]
11. Thivya, C.; Chidambaram, S.; Rao, M.S.; Thilagavathi, R.; Prasanna, M.V.; Manikandan, S. Assessment of
fluoride contaminations in groundwater of hard rock aquifers in Madurai district, Tamil Nadu (India). Appl.
Water Sci. 2017, 7, 1011–1023. [CrossRef]
12. Walia, T.; Abu Fanas, S.; Akbar, M.; Eddin, J.; Adnan, M. Estimation of fluoride concentration in drinking
water and common beverages in United Arab Emirates (UAE). Saudi Dent. J. 2017, 29, 117–122. [CrossRef]
13. Edmunds, W.M.; Smedley, P.L. Fluoride in natural waters. In Essentials of Medical Geology—Impacts of the
Natural Environment on Public Health; Selinus, O., Ed.; Elsevier Academic Press: Burlington, MA, USA, 2005.
14. United States Public Health Service (USPHS). U.S. Public health service recommendation for fluoride
concentration in drinking water for the prevention of dental caries. Rep. Recomm. 2015, 130, 318–331.
15. Guissouma, W.; Hakami, O.; Al-Rajab, A.J.; Tarhouni, J. Risk assessment of fluoride exposure in drinking
water of Tunisia. Chemosphere 2017, 177, 102–108. [CrossRef]
16. Samal, A.C.; Bhattacharya, P.; Mallick, A.; Ali, M.M.; Pyne, J.; Santra, S.C. A study to investigate fluoride
contamination and fluoride exposure dose assessment in lateritic zones of West Bengal, India. Environ. Sci.
Pollut. Res. Int. 2015, 22, 6220–6229. [CrossRef]
17. Huang, G.; Gu, X. Water environment status analysis and policy research of drinking water sources in
Yancheng. Zhihuai 2009, 12, 19–21. (In Chinese)
18. Chae, G.; Yun, S.; Mayer, B.; Kim, K.; Kim, S.; Kwon, J.; Kim, K.; Koh, Y. Fluorine geochemistry in bedrock
groundwater of South Korea. Sci. Total Environ. 2007, 385, 272–283. [CrossRef]
19. Salifu, A.; Petrusevski, B.; Ghebremichael, K.; Buamah, R.; Amy, G. Multivariate statistical analysis for
fluoride occurrence in groundwater in the Northern region of Ghana. J. Contam. Hydrol. 2012, 140–141, 34–44.
[CrossRef]
20. Sultana, J.; Farooqi, A.; Ali, U. Arsenic concentration variability, health risk assessment, and source
identification using multivariate analysis in selected villages of public water system, Lahore, Pakistan.
Environ. Monit. Assess. 2014, 186, 1241–1251. [CrossRef]
Int. J. Environ. Res. Public Health 2019, 16, 1030 16 of 16

21. US Environmental Protection Agency (EPA). Guidelines for developmental toxicity risk assessment. Fed.
Regist. 1991, 56, 63798–63826.
22. Gu, T.; Wang, C.; Lian, S.; Fan, W.; Bo, X.; Zhao, S.; Zhang, Y. Discussion of fluorosis area distribution and
controp measures in Jiangsu Province. Prev. Control Endem. Dis. China 1991, 6, 304. (In Chinese)
23. Karro, E.; Indermitte, E.; Saava, A.; Haamer, K.; Marandi, A. Fluoride occurrence in publicly supplied
drinking water in Estonia. Environ. Geol. 2006, 50, 389–396. [CrossRef]
24. Saxena, V.K.; Ahmed, S. Dissolution of fluoride in groundwater: A water-rock interaction study. Environ.
Geol. 2001, 40, 1084–1087.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).