Fluoride Exposure and the Effect of Tobacco Smoking on Urinary Fluoride Levels in Primary Aluminum Workers

: Workers in primary aluminum smelter are exposed to ﬂuoride from cryolite (Na 3 AlF 6 ) used in the electrolysis process. Post-shift urinary ﬂuoride is considered as an appropriate index for exam-ination of ﬂuoride exposure. The objective of the study was to investigate the exposure to ﬂuoride in primary aluminum smelter in Žiar nad Hronom (Slovakia) during three consecutive two-year periods between 2012 and 2018. The relationship between ﬂuoride exposure in the occupational environment, tobacco smoking, and pre- and post-shift urinary ﬂuoride concentration was investigated in 76 male workers in the ages from 21 to 60 years. Workers were monitored by personal ﬂuoride sampling equipment. Their urinary samples were collected prior to the start and at the end of an eight-hour shift. Fluoride content in urine samples was analyzed by potentiometric ion-selective electrode and expressed as weight ratio of ﬂuoride content to creatinine. The Mean ± SD particulate ﬂuoride concentration in occupational air was 0.966 ± 1.658 mg/m 3 and gas-phase ﬂuoride concentration was 0.327 ± 0.809 mg/m 3 . Mean urinary ﬂuoride concentration of all workers was signiﬁcantly higher ( p < 0.001) after the eight-hour shift. Smokers tended to have a higher post-shift mean urinary ﬂuoride concentration than non-smokers, but this difference was not statistically signiﬁcant ( p = 0.11). The difference between these two groups of workers was also not statistically signiﬁcant ( p = 0.62) before the shift. Therefore, according to results, smoking caused no statistically signiﬁcant difference in urinary ﬂuoride levels between the group of smokers and group of non-smokers in primary aluminum workers.


Introduction
Primary aluminum production is one of the most important industries in the world with a total global production of 5438 thousand metric tonnes [1]. The mechanism of the primary aluminum production is electrolytic reduction (Hall-Héroult process) of aluminum oxide-alumina (Al 2 O 3 ) dissolved in a molten fluoride electrolyte consisting of cryolite (Na 3 AlF 6 ) at a temperature of about 960 • C. Electrolytic reduction is a continuous process inside steel container cells coated with carbon. It gives pure liquid aluminum metal at the negative carbon electrode (cathode) and carbon dioxide (CO 2 ) at the positive carbon electrode (anode) [2]. The anode slowly reacts under intense heat, carbon dioxide, and other gaseous emissions such as polycyclic aromatic hydrocarbons (PAH) and hydrogen fluoride (HF). Cell lines are situated in the potroom. Most current aluminum smelters use pre-bake anode technologies and the older plants use the Søderberg process [2,3].
The purpose of the fluoride components in the electrolyte is to lower the melting point to operate the cells at a lower temperature. Aluminum fluoride (AlF 3 ) neutralizes 2,000,000 tonnes of aluminum were produced in this factory. Alumina, as a base material for the production of primary aluminum, was produced from 1957 from purchased bauxite. This technology, in view of the environment and occupational conditions, was unsuitable. Due to anode paste baking and insufficient exhaustion of fumes to a treatment facility, enormous volumes of tar and fluorine were emitted into the environment. Between 1992 and 1995, there was a complete replacement of electrolysis technology; the Søderberg cells were replaced by a pre-bake anode system. Pre-bake technology significantly reduced the fugitive emission in the occupational environment due to its hermetic construction. During the research period, the number of cells in operation increased from 175 to the current 226. Each of the cells is connected to the suction and cleaning system, while during anode replacement, the suction capacity is increased. There are also organizational measures in place that limit the current number of open cells at the same time.
In Žiar nad Hronom, fluoride intake by ambient air was significant until 1992 when 10-30% of all samples exceeded the highest limit value for fluoride in air (5.0 mg/m 3 ). Between 1992 and 1995, classic Søderberg cells were gradually replaced with new technology and the concentration of fluoride in ambient air decreased to <1.0 µg/m 3 [12].
It should be noted that the concentrations of fluoride in stack emissions from smelter met the legislative requirements represented by emission limits. The emission limit for organized waste air exhausts in Slovak legislation has been determined only for HF at the level of 2.0 mg/m 3 .
Further modernization of technological equipment and tightening of requirements for operational practice and emission limits (according to Best Available Techniques) continued to reduce the amount of fluoride emitted from the primary aluminum production plant in Žiar nad Hronom in following years.

Study Participants
In total, 76 male workers ranging in the ages from 21 to 60 years in an aluminum smelter were evaluated by spot urine samples. The age of each individual was not recorded for the purposes of the study. Data  Samples were collected prior to the start of an eight-hour shift and again at the end of the shift. Sampling occurred during three consecutive two-year periods between 2012 and 2018. Each study participant had been performing his working activity for more than a year. Work positions located directly in the potroom were investigated; in this area, fluoride exposure during the shift is at the highest level. At the same time, chemical determination and analysis of fluoride content in occupational air was carried out.
Samples of urine were analyzed for fluoride; results were expressed as weight ratio of fluoride content to creatinine (urinary fluoride). Workers' smoking habits were obtained using a questionnaire. The distribution of primary aluminum production workers who participated in the study was relatively uniform over the duration of the study. In total, 12 work positions were monitored (see Table 1). The worker position was not recorded in one instance. Of the 76 study participants, 35 were regular smokers (>10 cigarettes per day) and 39 were non-smokers. In two cases, the participants did not comment on the questionnaire about their smoking habits. Due to shift work, employee fluctuation, and the two-year study period (in accordance with the requirements of the legislation [16]), it was not possible to examine real working conditions every year for the same participants.

Number of Participants
Work Position n = 7 Crane operator n = 10 Forklift operator n = 2 Anode transport n = 1 Casting of alloys n = 8 Preparation of smelting in the foundry, transport of dross n = 11 Cell operator n = 19 Anode replacement n = 7 Pumping wagon operator n = 1 Primary refining n = 3 Cell start n = 1 Measurement technician n = 5 Pumping and transport of liquid metal For more detailed information on monitored participants in each year of the research, see Appendix A.
The majority (84%) of workers who participated in the study came either directly from the city of Žiar nad Hronom, Slovakia, or from the surrounding area (<30 km). Thus, it is possible that participants also had environmental exposure to fluoride through potable water, ambient air, or by edibles. The content of fluoride in ambient air was discussed in Section 2.2 and its impact as confounding factor was considered negligible in this research. Fluoride intake by potable water is minimal in Žiar nad Hronom because the city and its surroundings are supplied by bulk potable water from distant non-contaminated sources where the quality of potable water must meet the hygienic standards (limit value 1.50 mg fluoride per liter). This water is also used as drinking water by employees of the smelter, thus also by all study participants. According to the latest comprehensive study of fluoride occurrence in Žiar nad Hronom, fluoride content from the bulk potable water supply and from private potable water wells from 1986 to 1997 did not exceed the highest limit value for fluoride for potable water; the average values in individual localities ranged from <0.15 to 1.45 mg/L. It should be noted that a major change in electrolysis technology occurred in 1992-1995, when a new electrolysis was built; the old blocks of the plant were gradually disconnected [12].

Air Sampling and Analyses
Sampling of fluoride in the occupational air was performed in accordance with the NIOSH method, using nitrocellulose membrane filters with a pore size of 0.45 µm for particulate fluoride sampling [17]. Filters were placed inside plastic cassettes together with MILLIPORE cellulose filters impregnated with sodium formate for gaseous phase sampling. The sampling apparatus was supplemented by a calibrated pump and worn by each participant during the shift. After sampling, the exposed filters were dried in a desiccator. Subsequently, nitrocellulose membrane filters were decomposed in sodium hydroxide solution and the solid fluoride content was analyzed by potentiometric ion-selective electrode on an ORION 4 STAR analytical instrument. Gas-phase fluoride captured on impregnated cellulose filters were analyzed using the same procedure. The analytical method was controlled by certified reference material ASTASOL. The determination of the particulate concentration was performed by the gravimetric method. The weight of the collected fluoride content was determined from the difference in weight of the filters before and after the measurement. Particulate concentration was calculated by dividing the weight of particulate matter collected on the filter by the pumped air volume recalculated on standard condition (T = 293.15 K, p = 101.325 kPa).
Workers in the potroom wore personal respiratory protection equipped with organic vapors/acid gas cartridges and dust prefilter.

Urine Sampling and Analyses
The urine samples of each participant were taken before and after an eight-hour shift. The urine was collected in sterile plastic containers. The analysis of fluoride in urine was performed by the NIOSH method [18]. In the analysis, the urine was diluted with an auxiliary solution of Total Ionic Strength Adjustment Buffer (TISAB) to adjust the pH and ionic strength of the solution to release fluoride ions from possible aluminum and ferric complexes. TISAB was prepared by diluting 57 mL of glacial acetic acid, 58 g sodium chloride, and 4 g trans-1,2-diaminocyclohexane-N,N,N ,N -tetraacetic acid monohydrate in 1000 mL water and subsequently adjusted to pH 5.2-5.4 with sodium hydroxide, 20% (w/v). A sample of 25 mL of urine was added to an equal volume of auxiliary solution (TISAB). Subsequently, magnetic stirring was switched on. Two electrodes were used for the determination: ion selective fluoride electrode and a calomel electrode, both immersed in stirring solution. It is necessary to wait until the potential stabilizes while stirring (approximately 10 min). The fluoride concentration in urine was acquired from the calibration curve.
Picric acid reacts with creatinine in an alkaline medium to produce an orange-red color, which is measured photometrically. The determination of creatinine concentration in the urine is carried out in order to suppress errors in the calculation of the foreign substance concentration due to diuresis. Creatinine is excreted from an organism at a relatively constant rate, so the data calculated for creatinine are more accurate than the concentration of native urine [19]. The result is expressed as the amount of fluoride in mg corresponding to 1 g of creatinine. The determination of creatinine in urine was performed by NIOSH methods [20,21].

Statistical Analysis
Data were analyzed by STATISTICA (StatSoft Inc., Tulsa, OK, USA, ver. 12). Statistical significance was set at p < 0.05. Normality of the data was evaluated by Shapiro-Wilk (S-W) test and Kolmogorov-Smirnov (K-S) test. The results of these tests rejected normality of all variables (p < 0.05). Table 2 shows the descriptive statistics for all study participants (for each study participant s data, see Appendix A), group of smokers and non-smokers including measured data of air samples and urinary fluoride samples.

Results and Discussion
Among the monitored work positions in Table 1 there were no statistically significant differences (p > 0.05) for measured fluoride concentrations (particulate and gas-phase fluoride) in the occupational air. Within the total fluoride exposure, workers are the most exposed to the particulate fluoride emissions. The highest total (particulate + gas phase) fluoride concentration (12.120 mg/m 3 ) was registered in anode replacement work position. Our findings are consistent with the assumption that during anode replacement the worker is temporarily exposed to the open environment of the steel container and despite the increased efficiency of the air conditioning, the highest rate of fluoride exposure occurs in this operation. By contrast, the lowest concentration (0.018 mg/m 3 ) was found for the crane operator who is relatively isolated from direct exposure in the crane cabin.
A similar conclusion was reached by Ehrnebo and Ekstrand [22] who investigated 41 workers in an aluminum plant in Sweden during an eight-hour work shift; the mean total fluoride exposure in this study was 0.91 mg/m 3 , of which 34% was HF. For comparison, Seixas et al. [9] found mean particulate fluoride and gas-phase fluoride (HF) concentrations were 4.1 and 0.7 mg/m 3 when they evaluated the relationship between particulate fluoride exposure and urinary excretion in an aluminum smelter. Contrary to determined distribution of particulate and gas-phase fluoride in the occupational environment of investigated potroom, the Risk assessment report [4] concluded that the average fluoride emission of a primary aluminum smelter within the EU consists of 53% HF and 47% particulate fluoride. This may be explained due to the composition of study participants. The majority of workers were exposed to emissions during anode replacement in cells, including the particulates during crust cleaning of the anode cavity before installing new anodes. Gas-phase fluoride is formed in electrolysis process directly in the cell and it is emitted especially during anode replacement process; however, particulates are formed from handling by application of cryolite powder in the molten bath mixture and when crust is disrupted. Crust is a solid matter transformed from anode covering material by heat [23].
The Kruskal-Wallis test (K-W test) showed that there was no statistically significant difference (p = 0.57) in particulate fluoride concentrations in the monitored years of the research. However, gas-phase concentrations were statistically significantly different (p = 0.003). As can be seen in Figure 1, this was due to increased gas-phase fluoride concentrations in 2014 (an approximately five times higher mean value compared to other years).
Occupational Safety and Health Administration (OSHA) permissible exposure limit (PEL) of fluorides (as F) at a value of 2.5 mg/m 3 (as time-weighted average) was exceeded (2.67-10.29) in 10 workers (see Appendix A) [24]. Compared to the general population, inhalation of fluoride present in ambient air does not contribute more than 0.01 mg per day to the total fluoride intake [25]. Since 1996, long-term average annual concentrations of fluorides in ambient air are <1 µg/m 3 in the district of Žiar nad Hronom [12,26]. Figure 2 shows a comparison of mean ± SE urinary fluoride concentrations during the research period. No statistically significant difference (p > 0.05) was found for either pre-or post-shift urinary fluoride over the monitored years.
The most effective way to examine fluoride exposure provides the relation between the concentration in the air and the quantities excreted in the urine at the end of the shift [27].
The results of correlation analysis between fluoride exposure in the occupational environment (i.e., by inhalation) and corresponding urinary fluoride levels in monitored participants are presented in Table 3 and Figure 3. Occupational Safety and Health Administration (OSHA) permissible exposure limit (PEL) of fluorides (as F) at a value of 2.5 mg/m 3 (as time-weighted average) was exceeded (2.67-10.29) in 10 workers (see Appendix A) [24]. Compared to the general population, inhalation of fluoride present in ambient air does not contribute more than 0.01 mg per day to the total fluoride intake [25]. Since 1996, long-term average annual concentrations of fluorides in ambient air are <1 µg/m 3 in the district of Žiar nad Hronom [12,26]. Figure 2 shows a comparison of mean ± SE urinary fluoride concentrations during the research period. No statistically significant difference (p > 0.05) was found for either pre-or post-shift urinary fluoride over the monitored years.   Occupational Safety and Health Administration (OSHA) permissible exposure limit (PEL) of fluorides (as F) at a value of 2.5 mg/m 3 (as time-weighted average) was exceeded (2.67-10.29) in 10 workers (see Appendix A) [24]. Compared to the general population, inhalation of fluoride present in ambient air does not contribute more than 0.01 mg per day to the total fluoride intake [25]. Since 1996, long-term average annual concentrations of fluorides in ambient air are <1 µg/m 3 in the district of Žiar nad Hronom [12,26]. Figure 2 shows a comparison of mean ± SE urinary fluoride concentrations during the research period. No statistically significant difference (p > 0.05) was found for either pre-or post-shift urinary fluoride over the monitored years.    No significant correlation between fluoride concentrations in the occupational environment and in urine corrected to creatinine was found in the present study. Pierre et al. [27] reported an association between fluoride concentration of 2.5 mg/m 3 (OSHA PEL) in an occupational environment and post-shift urinary fluoride concentration of 6.4 mg/g creatinine, and a peak value of 7.4 mg/g creatinine, respectively. As mentioned above, our results showed OSHA PEL exceeded in 10 workers. In contrast with Pierre et al. [27], postshift urinary fluoride levels in this workers reached only 0.2-1.2 mg/g creatinine and peak post-shift urinary fluoride 3.73 mg/g creatinine corresponds to 0.16 mg/m 3 of total fluoride concentration in the occupational environment. More than 50% exceeded samples above OSHA PEL were associated with anode handling working position (i.e., anode replacement, cell operator). Post-shift urinary fluoride levels in these workers proved that personal respiratory protection used during high exposure periods while working over open cells was effective. A similar conclusion was reached by Seixas et al. [9] who ascribed the No significant correlation between fluoride concentrations in the occupational environment and in urine corrected to creatinine was found in the present study. Pierre et al. [27] reported an association between fluoride concentration of 2.5 mg/m 3 (OSHA PEL) in an occupational environment and post-shift urinary fluoride concentration of 6.4 mg/g creatinine, and a peak value of 7.4 mg/g creatinine, respectively. As mentioned above, our results showed OSHA PEL exceeded in 10 workers. In contrast with Pierre et al. [27], post-shift urinary fluoride levels in this workers reached only 0.2-1.2 mg/g creatinine and peak post-shift urinary fluoride 3.73 mg/g creatinine corresponds to 0.16 mg/m 3 of total fluoride concentration in the occupational environment. More than 50% exceeded samples above OSHA PEL were associated with anode handling working position (i.e., anode replacement, cell operator). Post-shift urinary fluoride levels in these workers proved that personal respiratory protection used during high exposure periods while working over open cells was effective. A similar conclusion was reached by Seixas et al. [9] who ascribed the lack of relationship between particulate fluoride exposure and the post-shift urinary fluoride by the effective use of respiratory protection in carbon setters (anode replacement).
Kono et al. [28] found a linear relationship between HF concentration (>5 ppm) in the occupational environment and mean urinary fluoride levels in electronics industry workers. The wide variation of fluoride levels in urine has been ascribed to the consumption of tea and seafood (water intake was not considered because only <1% of the water supply is fluoridated in Japan). In agreement with Kono et al. [28], we also rejected the fluoride intake by potable water in Žiar nad Hronom which met hygienic standards. In the occupational environment, fluoride intake via inhalation can reach 16.8 mg per day; nevertheless, daily intake of fluoride by optimally fluoridated potable water is about 1.4-2.4 mg [8]. During our research, participants were asked to avoid tea consumption and use of toothpastes enriched with fluoride for at least 48 h prior to urine collection; however, an apparent limitation of this procedure is its verification. Waugh et al. [29] and Koç et al. [13] concluded that tea is a significant source of fluoride intake. A major source of limitation in this study is fluoride intake effected by the composition of diet. With low water fluoride, the urinary fluoride concentration is much more influenced by eating habits [27]. According to Whitford [30], vegetarian diet leads to more alkaline urine and in addition increases the urinary fluoride excretion. Assessment of fluoride intake by food is quite challenging due to exposure to multiple dietary sources and supplements. Some authors in other studies used diet standardization to reduce this limitation, however, it is quite difficult to conduct [7].
Urinary fluoride concentrations of 0.8-1.2 mg/L are regarded as indicating optimal exposure to fluoride in population [31]. Figure 4 compares pre-and post-shift urinary fluoride concentrations for all study participants.
toothpastes enriched with fluoride for at least 48 h prior to urine collection; however, an apparent limitation of this procedure is its verification. Waugh et al. [29] and Koç et al. [13] concluded that tea is a significant source of fluoride intake. A major source of limitation in this study is fluoride intake effected by the composition of diet. With low water fluoride, the urinary fluoride concentration is much more influenced by eating habits [27]. According to Whitford [30], vegetarian diet leads to more alkaline urine and in addition increases the urinary fluoride excretion. Assessment of fluoride intake by food is quite challenging due to exposure to multiple dietary sources and supplements. Some authors in other studies used diet standardization to reduce this limitation, however, it is quite difficult to conduct [7].
Urinary fluoride concentrations of 0.8-1.2 mg/L are regarded as indicating optimal exposure to fluoride in population [31]. Figure 4 compares pre-and post-shift urinary fluoride concentrations for all study participants. According to Wilcoxon signed-rank test, mean post-shift urinary fluoride concentration was significantly (p < 0.001) higher compared to mean pre-shift urinary fluoride concentration for all investigated participants. From this, it can be concluded that there is an According to Wilcoxon signed-rank test, mean post-shift urinary fluoride concentration was significantly (p < 0.001) higher compared to mean pre-shift urinary fluoride concentration for all investigated participants. From this, it can be concluded that there is an increase in fluoride concentration in primary aluminum workers during their work performance, although the relationship between occupational exposure and urinary fluoride levels was not confirmed. Overall, these findings are in accordance with previous findings reported by Rees et al. [32] who found significant (p < 0.05) difference between mean pre-and post-shift concentrations in workers exposed above OSHA PEL (2.5 mg/m 3 ) to calcium fluoride (CaF 2 ), and a weak correlation between intensity of occupational exposure to calcium fluoride and post-shift urinary fluoride concentration. Mean fluoride exposure in the dustiest environment was 24.3 mg/m 3 (40.6% respirable fraction) which is approximately 10 times OSHA PEL. Only one urinary concentration exceeded the recommended limit value of 7 mg fluoride per liter. It is important to highlight the fact that potroom of aluminum smelter is a high-temperature workplace; therefore, the relating of the urinary excretion of fluoride to g-creatinine is necessary. According to the World Health Organization (WHO) [31], the mean 24-h urinary creatinine value is 15 mg/kg of body weight per day. Several previous studies have not reported creatinine correction. Zohouri et al. [33] reported the mean urinary fluoride concentration of 1.49 (±0.63) mg/g creatinine in children.
The biological exposure index (BEI) for fluoride exposure is 3 mg/g creatinine prior to shift and 10 mg/g creatinine at end of shift [34]. The BEI were not exceeded in the monitored spot urine samples. We found that maximum pre-and post-shift urinary fluoride concentrations were 2.42 mg/g creatinine and 3.73 mg/g creatinine.
In agreement with WHO [31], the time of urine collection was approximately equal during the whole study period, which is important for correct interpretation of the results, because urine that has accumulated in the body over a shorter time period may manifest a short-lived peak of the fluoride concentration. The findings of the present study demonstrate variations in urinary fluoride levels within participants. We assume that improper use of a personal respiratory protection might contribute to these differences.
According to Seixas et al. [9], mean urinary fluoride concentrations were 1.3 and 3.0 mg/g creatinine in pre-shift and post-shift urine samples. The results of our study showed markedly lower mean concentrations of particulate fluoride and HF in the occupational environment. Our results also showed lower pre-and post-shift urinary fluoride concentrations. Research by Seixas et al. [9] was performed on a sample of 32 workers in a shorter period (days 1 and 3 of a three-day workweek), making it more sensitive to local and short-term increases in fluoride concentration in the occupational environment. On the other hand, our study provided information obtained from a longer time period and various operational situations that may have occurred in aluminum smelter.
A frequency distribution in Figure 5 shows that 89% of all study participants had urinary fluoride difference (post-shift minus pre-shift concentration) below 1 mg/g creatinine, with the maximum value at 3.38 mg/g creatinine and the minimum was at a negative value of −0.82 mg/g creatinine (a negative result of urinary fluoride difference was found in nine workers). According to Lauwerys et al. [35], the difference between post-and pre-shift urinary fluoride concentration should not exceed 3 to 4 mg/g creatinine, respectively, pre-shift values should be <3 mg/g creatinine. Tobacco smoking was the main confounding factor of fluoride exposure we decided to investigate. In total, 35 study participants were regular smokers consuming at least 10 cigarettes per day. Other 39 participants declared they were strict non-smokers and they were not aware of the regular secondhand smoke exposure. There was a separate smoking area in smelter. Workers had the right to one uninterrupted 30-min rest break during their working day, if they worked 8 h per day. Majority of smokers took this break also as smoking break. Further, each of the smokers confirmed that they smoked at least one cigarette before their shift. Figure 6 shows a comparison of mean pre-and post-shift urinary fluoride concentrations according to smoking habits of investigated primary aluminum workers. The results in the present study showed no statistically significant difference of the urinary fluoride level among group of smokers and non-smokers before the shift (p = 0.62) and after the shift (p = 0.11) by K-W test. This suggests that under the given research conditions, tobacco Tobacco smoking was the main confounding factor of fluoride exposure we decided to investigate. In total, 35 study participants were regular smokers consuming at least 10 cigarettes per day. Other 39 participants declared they were strict non-smokers and they were not aware of the regular secondhand smoke exposure. There was a separate smoking area in smelter. Workers had the right to one uninterrupted 30-min rest break during their working day, if they worked 8 h per day. Majority of smokers took this break also as smoking break. Further, each of the smokers confirmed that they smoked at least one cigarette before their shift. Figure 6 shows a comparison of mean pre-and post-shift urinary fluoride concentrations according to smoking habits of investigated primary aluminum workers. The results in the present study showed no statistically significant difference of the urinary fluoride level among group of smokers and non-smokers before the shift (p = 0.62) and after the shift (p = 0.11) by K-W test. This suggests that under the given research conditions, tobacco smoking does not affect the fluoride content of workers in primary aluminum smelter. It remains unclear why non-smokers show higher or equal pre-shift mean urinary fluoride concentration (0.469 ± 0.379 mg/g creatinine) than smokers (0.441 ± 0.338 mg/g creatinine). We assume that pre-shift samples reflected the worker s total burden of fluoride and although smokers consumed a cigarette before shift, this was not reflected despite to rapid kinetics and the biological half-life of fluoride in human body. From the difference between pre-and post-shift urinary fluoride values, a decreasing trend of urinary fluoride concentration in smokers was found (r = −0.133, p = 0.447). The group of non-smokers was characterized by a slightly increasing trend (r = 0.03, p = 0.868).
Appl. Sci. 2021, 11, x FOR PEER REVIEW 12 of 17 and inhalable dust below OSHA PEL. A similar conclusion also reached Tu et al. [39], who concluded that age, work history, smoking, and alcohol consumption did not affect the blood fluoride and urinary fluoride levels in 300 workers from an aluminum plant in China. According to Koç et al. [13], urinary fluoride concentration is higher in smokers compared to non-smokers among 300 students at the University of Kafkas, Turkey. The present study found no significant differences between pre-and post-shift urinary fluoride levels in groups of smokers and non-smokers, although non-smokers had slightly higher levels of pre-shift urinary fluoride. This effect may be caused by other confounding variables, which we did not consider, such as drinking tea, dietary intake of fluoride, or metabolic disorders. Therefore, further research needs to be supplemented with data on excessive intake of fluoride.

Conclusions
The results of the present study showed a statistically significant increase of fluoride in the urine of primary aluminum workers during an eight-hour shift. The research procedure and methodology provide an effective indication of fluoride exposure in aluminum smelter in comparison with BEI. However, the source of fluoride intake in these workers was not clarified. The degree of correlation between occupational exposure to fluoride and urinary fluoride levels was ascribed to person respiratory protection. Future investigations are necessary to validate conclusions that can be drawn from this study. It will be important for future research to conduct repeated measurements for the same workers for a longer time period. A single spot urine sample from an individual may be The research on the interaction between tobacco smoking and fluoride absorption and, subsequently, excretion is limited. It is estimated that heavy cigarette smoking could daily contribute about 0.01 mg of fluoride intake per kg of body weight [36]. Laisalmi et al. [37] reported the relationship between smoking and inorganic fluoride levels due to anesthesia with enflurane (2-chloro-1,1,2,-trifluoroethyl-difluoromethyl ether). In contrast with the present study, they observed that the serum fluoride concentrations were significantly different between group of smokers and non-smokers. On the other hand, our results are in agreement with Radon et al. [38], who found no combined effect of smoking and occupational exposure in aluminum potroom workers exposed to hydrogen fluoride and inhalable dust below OSHA PEL. A similar conclusion also reached Tu et al. [39], who concluded that age, work history, smoking, and alcohol consumption did not affect the blood fluoride and urinary fluoride levels in 300 workers from an aluminum plant in China.
According to Koç et al. [13], urinary fluoride concentration is higher in smokers compared to non-smokers among 300 students at the University of Kafkas, Turkey. The present study found no significant differences between pre-and post-shift urinary fluoride levels in groups of smokers and non-smokers, although non-smokers had slightly higher levels of pre-shift urinary fluoride. This effect may be caused by other confounding variables, which we did not consider, such as drinking tea, dietary intake of fluoride, or metabolic disorders. Therefore, further research needs to be supplemented with data on excessive intake of fluoride.

Conclusions
The results of the present study showed a statistically significant increase of fluoride in the urine of primary aluminum workers during an eight-hour shift. The research procedure and methodology provide an effective indication of fluoride exposure in aluminum smelter in comparison with BEI. However, the source of fluoride intake in these workers was not clarified. The degree of correlation between occupational exposure to fluoride and urinary fluoride levels was ascribed to person respiratory protection. Future investigations are necessary to validate conclusions that can be drawn from this study. It will be important for future research to conduct repeated measurements for the same workers for a longer time period. A single spot urine sample from an individual may be affected by many strongly variable confounding factors. Collecting the post-shift urine samples after several working days would be appropriate and more representative. The study showed that tobacco smoking caused no statistically significant difference in urinary fluoride levels between group of smoker and non-smokers; however, further investigations are needed to confirm these results. Overall, further research should address these issues. We would like to focus on the assessment of fluoride intake from tea consumption, fluoride toothpastes, and dietary intake in primary aluminum workers.  Institutional Review Board Statement: For this type of study a formal IRB approval is not required. Study involves data collected from individual subjects that cannot be identified directly or through identifiers linked to the subjects.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available within the article.

Acknowledgments:
The paper is based on work performed under research contract KEGA 008TU Z-4/2019. The authors gratefully thank Slovalco a.s. and PRO CARE Žiar nad Hronom for their assistance and cooperation.

Conflicts of Interest:
The authors declare that there is no conflict of interest. Table A1. Data on study participants and their occupational exposure (particulate and gas-phase fluoride concentration detected by personal sampling), urinary fluoride biological monitoring (pre-and post-shift urinary fluoride), tobacco smoking habit, and work position.