Determination of fluoride content in teas and herbal products popular in Poland

Abstract

Purpose

Fluoride level, due to its narrow therapeutical range, must be constantly monitored in beverages, especially in daily-consumed plant infusions. Fluoride is important for prevention of tooth decay and osteoporosis, but its excess leads to fluorosis. Since tea can selectively absorb fluorides from soils, the question arises if a long-term consumption can pose an adverse effect on human health.

Methods

Infusions of 33 popular teas (black, green, white, earl grey, pu-erh), tea-like products (rooibos, yerba mate) and herbs (chamomile, mint, nettle, purges, yarrow) available in the Polish market were analyzed with respect to a fluoride level by means of a validated ion-selective electrode method, which is proven to be fast and reliable.

Results

Significantly different fluoride concentrations in infusions were observed, with black tea on top, where extraction of fluoride is the highest (average 2.65 mg F⁻/L, range 0.718–6.029 mg/L). Two-fold higher fluoride contents were measured in infusions made from black tea bags than from leaves (average 3.398 mg/L and 1.529 mg/L, respectively). Green teas released comparable amounts of fluoride as black teas, while in herbal extracts the fluoride content was negligible.

Conclusions

The rank with respect to the fluoride concentration in an infusion is as follows: black tea > green tea > earl grey > pu-erh > white tea>>>rooibos, yerba mate, herbal products. Increasing of brewing time results in an increased fluoride content, but the overall content of fluoride in the analyzed infusions of teas and herbs was not high enough to cause a risk of fluorosis, even if left to brew up to 15 min.

Introduction

Fluorine is a reactive halogen element often present in nature, hence also in environmental samples, food and materials of biological origin. Fluoride (F⁻), in a suitable dose, is considered beneficial for humans, as it is necessary for growth of tissues, bones and teeth. An increased dosage, however, can cause damage to a human body, not only in the form of dental or skeletal fluorosis, but also to soft tissues, such as brain, kidney, liver and spinal cord [1]. The lethal dose for adults is from 0.20 to 0.35 g F⁻/kg of body weight [2]. According to WHO [3], fluoride is classified, next to arsenic and nitrate, as a contaminant in drinking water, which may cause large-scale health problems. This situation occurs mainly in endemic areas where drinking water sources have a high fluoride content, such as in parts of India, Africa, China, and Iran [4–7].

Fluoride can be released into the groundwater by slow dissolution of fluorine-containing rocks [8], but it may also leach out from anthropogenic sources such as, coal fired power stations, brick and iron works, beryllium extraction plans and aluminium smelters [9]. Moreover, in certain countries (e.g. the United States), drinking water is fluoridated (by adding sodium fluoride, sodium fluorosilicate and/or hydrofluorosilicic acid), which keeps a fluoride level at 0.7 mg/L [10, 11]. The amount of fluoride consumed with drinking water contributes vastly to fluoride intake. According to Malin and Till [11], before 2010 drinking water in US was even more fluoridated, reaching 1.2 mg/L, which led to serious fluoride exposure of children. In Poland water is not fluoridated, with the fluoride level below 0.3 mg/L (in the majority of the settlements in Poland), with just a few spots, where the concentration exceeds 2 mg/L [12]. Even without purposely fluoridated water, there is a need for monitoring fluoride in food and beverages. Many epidemiological observations showed that fluorosis may be caused by long-term consumption of large amounts of tea, such as observed in Tibet, Mongolia and in the western parts of China [13] Which concerned both adults and children [14]. The studies on a correlation between an elevated level of fluoride in drinking water and incidence of hypertension [15], infertility and risk of spontaneous abortion [16, 17], and Attention-Deficit Hyperactivity Disorder (ADHD) are particularly disturbing [11]. Considering that tea is the most popular non-alcoholic beverage in the world (two thirds of the world’s population drinks tea every day), there is no doubt that tea consumers must become aware about fluoride and its presence in tea infusions [18].

Among the plants, tea (Camellia sinensis L.) is described as „hyper-accumulator” of fluoride [19] because of selective absorption from soils, mainly acidic, rich with fluorine-containing minerals [20]. Historically, tea was grown only under natural conditions, but nowadays many tea growers apply phosphate fertilizers [10] or fluorine-containing pesticides [21] which would increase a fluoride content in tea plants even more. It should also be emphasized that absorption and accumulation of fluoride is proportional to the concentration in the soil [22] and progresses in time, which makes aged leaves much richer with fluoride [20]. Gao et al. [23] revealed, that fluoride in a tea plant is accumulated in a soluble fraction of tea roots and leaves. It is located mainly in vacuoles (98.1% of total fluorides measured in the protoplasts), which reduces its toxicity to organelles, hence leaves with a high content of fluoride did not indicate any symptoms of toxicity and did not differ from the others. During a brewing process fluoride contained in tea leaves is released into an infusion; creating a potential health risk related to the consumption of tea for adults and children [24]. According to Fung et al. [25] 53–80% of total fluoride accumulated in tea leaves can be released, but it depends quite significantly on a type of brewed tea.

The history of tea goes back more than 5000 years in Asian countries, such as China, Japan, India and Thailand [26]. As reported on 23th session of FAO Intergovernmental Group on Tea, production reached 5.73 million tons in 2016 year (5.35 million tons in 2013 year). Although China, being a cradle of tea, is still a leader of its production (2.44 million tons in 2016), nowadays tea is produced in 45 countries. Countries with an old tradition and culture of drinking tea are major tea producers, i.e. India, Kenya, Sri Lanka, Vietnam, Turkey and Iran [27]. Poland, on the other hand, is a major importer of tea, reaching about 38 thousand tons in 2019 [28].

Drinking tea can show health-promoting effects, related to a presence of tannins (which are able to bind toxins), polyphenols (cholesterol decreasing, antioxidant, antimutagenic and anticancer properties) [29], theanine, caffeine (stimulant properties), and bioactivities of catechins or polysaccharides [30]. However, it can also have a negative effect, not only because of fluoride. Tea may contain harmful elements such as Al [18, 31], As, Cr, Cd, Cu, Fe, Mn, Ni and Pb [18, 32, 33], as well as organic contaminants, like pesticide residues or mycotoxins [34–36] and process-induced toxicants as acrylamide [37, 38] and polycyclic aromatic hydrocarbons [39].

We aimed our research at determination of a fluoride content in infusions of popular teas in Poland, accessible in local supermarkets. Our research covered various types and brands of tea, as well as few tea-like products used for making beverages called “tea”, but produced from other plants, such as rooibos (Aspalathus linearis) or yerba mate (Ilex paraguariensis), consumed broadly for their alleged health-beneficial properties or simply for their taste. We also included the most popular herbs, which are willingly consumed in Poland, brewed similarly to tea. To measure fluoride concentration an ion-selective fluoride electrode (fluoride-ISE) was used. The screening of fluoride content in the collection of teas and tea-like products is often presented in scientific literature, particularly in the countries wherein tea consumption is really high, such as: China [40], Denmark [41], Iran [42] and [43], UK [44] and [45], Japan [46], New Zealand [47], Sri Lanka [48], Ireland [49], etc. The reasoning behind this research is always a need to monitor a potential health risk due to fluoride exposure. In our work we verified which brand/type of tea and tea-like products (including herbs) releases a higher level of fluoride and how it is affected by a brewing time. Ultimately, evaluation of average tea consumption and potential risk of fluoride overdosing was presented.

Materials and methods

Sample collection and preparation

Twenty-four tea samples, four tea-like products (rooibos and yerba mate) and five herbal products were evaluated in this study. All products were purchased from supermarkets in Poland (Table 3). Before preparing the infusions, the samples were stored under dry conditions at room temperature.

Table 3.

Content of fluorides [mg/L] in infusions of selected teas and herbs (replicates n = 9, mean ± SD)

Brand	Package type	Origin	Brewing time [min]
			5		15
			Content ± SD [mg/L]	RSD [%]	Content ± SD [mg/L]	RSD [%]
black tea	bag	China	3.172 ± 0.066	2.08	4.085 ± 0.060	1.47
black tea	bag	India	2.212 ± 0.015	0.66	2.540 ± 0.012	0.77
black tea	bag	Sri Lanka	2.178 ± 0.051	2.35	2.442 ± 0.046	1.90
black tea	bag	India, Indonesia	4.480 ± 0.034	0.75	6.098 ± 0.149	2.45
black tea	bag	India, Africa	6.029 ± 0.127	2.10	6.759 ± 0.143	2.12
black tea	leaf	Turkey	1.564 ± 0.027	1.72	2.108 ± 0.104	4.93
black tea	leaf	Turkey	1.836 ± 0.032	1.75	2.088 ± 0.083	3.96
black tea	leaf	Turkey	1.595 ± 0.060	3.73	1.890 ± 0.045	2.39
black tea	bag	Sri Lanka	2.320 ± 0.013	0.56	2.948 ± 0.064	2.15
black tea	leaf	Sri Lanka	1.122 ± 0.032	2.87	1.376 ± 0.052	3.79
green tea	bag	NA	4.893 ± 0.053	1.08	4.908 ± 0.167	3.40
green tea	bag	NA	4.041 ± 0.108	2.68	4.494 ± 0.125	2.78
green tea	leaf	NA	2.051 ± 0.042	2.06	2.384 ± 0.102	4.27
green tea	leaf	China, India, Sri Lanka	0.832 ± 0.030	3.57	0.997 ± 0.034	3.40
green tea	leaf	NA	0.869 ± 0.018	2.10	0.944 ± 0.045	4.72
white tea	bag	China	2.258 ± 0.026	1.14	2.476 ± 0.120	4.85
white tea	bag	NA	2.359 ± 0.039	1.66	2.492 ± 0.010	0.38
white tea	bag	India, Africa	1.122 ± 0.006	0.51	1.186 ± 0.047	3.99
earl grey	bag	Asia	4.070 ± 0.054	1.32	4.722 ± 0.095	2.01
earl grey	bag	Sri Lanka	2.290 ± 0.021	0.92	2.930 ± 0.062	2.11
earl grey	leaf	NA	0.718 ± 0.012	1.66	0.817 ± 0.028	3.45
pu-erh	bag	China	2.634 ± 0.059	2.24	2.673 ± 0.096	3.57
pu-erh	bag	NA	2.169 ± 0.019	0.88	2.351 ± 0.091	3.88
pu-erh	leaf	China	1.186 ± 0.003	0.28	1.364 ± 0.023	1.65
rooibos (Aspalathus linearis)	bag	RPA	0.090 ± 0.004	4.47	0.091 ± 0.004	4.73
rooibos (Aspalathus linearis)	bag	RPA	0.038 ± 0.002	4.36	0.053 ± 0.002	3.81
yerba mate (Ilex paraguariensis)	leaf	Argentina	0.045 ± 0.002	4.31	0.066 ± 0.003	3.84
yerba mate (Ilex paraguariensis)	fine leaf	NA	0.041 ± 0.002	4.14	0.057 ± 0.003	4.53
chamomile (Chamomillae anthodium)	bag	Poland	0.043 ± 0.002	4.95	0.056 ± 0.002	3.18
mint (Menthae piperitae folium)	bag	Poland	0.077 ± 0.001	1.60	0.086 ± 0.004	4.83
nettle (Urticae folium)	bag	Poland	0.059 ± 0.003	4.55	0.075 ± 0.003	4.56
purges (Cistus incanus)	herbal drought	Albania	0.044 ± 0.002	4.46	0.083 ± 0.004	4.41
yarrow (Achillea millefolium)	bag	Poland	0.063 ± 0.005	7.80	0.065 ± 0.003	3.97

NA- data not available

The protocol of preparing an infusion (1% w/v water extract) was adjusted to mimic a domestic manner of brewing, i.e. one tea bag (selected randomly from each branded box of tea) or one teaspoon (sample weight 2.000 g) per one cup of boiling water (200 mL). The tea or herbal product was poured with 200 mL of deionized water (conductivity 0.08 µS/cm) at the temperature of 100^oC [20, 50].

The brewing process was conducted in a glass beaker covered with a lid. After extraction, the tea was removed and the infusion was cooled down to a room temperature (18–20^oC). To verify an influence of brewing time, we applied two time spans: 5 and 15 min. After brewing, the infusions were filtered through a cellulose filter. The samples were prepared in triplicates.

Fluoride determination

The analysis of fluoride (F⁻) level was made according to EPA Method 9214 by a combined fluoride ion-selective electrode (Orion™ Fluoride Electrode 9609, Thermo Scientific, USA), containing a monocrystalline form of lanthanum fluoride (LaF₃) as an ion-sensitive membrane and Ag/AgCl system as a reference electrode. The measurements were performed with an ion-meter/pH-meter CPI-505 (Elmetron, Poland) in triplicates. To stabilize an ionic strength and pH and to remove interfering ions, the infusions were spiked with TISAB (Total Ionic Strength Adjustment Buffer) in 1:1 ratio. TISAB, due to a presence of CDTA reagent, ensured the preferentially complexation of interfering aluminum and iron (III) cations up to the level of 5 ppm. TISAB was prepared of 57 mL of glacial acetic acid, 56 g of sodium chloride, 136 g of sodium acetate, and 4 g of CDTA (trans 1,2-diamino-cyclohexane-N,N,N’,N’-tetraacetic acid monohydrate) dissolved in 500 mL of deionized water, adjusted to pH about 5-5.5 with 5 M NaOH solution, and diluted up to 1 L [Method 9214]. The measurements were carried out until the stable potential value was observed, but not longer than five minutes after immersing the electrode in the solution.

The calibration was done in the range from 0.1 to 30 mg/L of fluoride at six calibration levels (EPA Method 9214). An intermediate standard solution of fluoride with a concentration 100 mg/L was prepared by diluting a stock solution (1000 mg/L, Sigma Aldrich, USA). Fluoride standard solutions of lower concentrations were obtained through further dilution. All solutions were prepared with deionized water II class according to EN ISO 3696 − 1999 (Hydrolab System HLP10p, Poland). They were stored in polyethylene containers in 8°C. The deionized water spiked with TISAB was taken as a blank sample.

The samples with fluoride standard addition (1.0 mg/L) were measured to perform method validation. Certified Reference Material (CRM) QC3162 (traceable to: NIST SRM 3183) Fluoride in Water (Sigma Aldrich, USA) with a certified content of fluoride 0.420 ± 0.0147 mg/L was also used for assessment of method accuracy.

Statistical analysis

All results were expressed as the mean ± standard deviation (SD). The data were analyzed by one-way ANOVA using Statgraphics Plus 3.0. The differences among treatments were determined by comparison of means using Tukey test. The level of statistical significance was considered at p<0.05.

Results and discussion

Method validation

Linearity

The calibration was performed using a series of standards with the defined fluoride concentrations: 0.1, 0.3, 1.0, 3.0, 10.0 and 30.0 mg/L. The calibration curve reflects a linear relationship between a fluoride concentration in a standard solution and mV-reading for this standard. The standards were run in triplicates, and the mean values were used for the calibration curve (Fig. 1). The linear regression coefficient R² was 0.9998.

Fig. 1.

Calibration curve in the range 0 ÷ 30.0 mg/L fluoride in water (n = 3)

Limit of detection (LOD) and limit of quantification (LOQ)

LOD and LOQ were calculated according to the equations, usually applied to ion-selective measurements [51]:

$$\begin{array}{c}\mathrm{LOD}=3*\mathrm{SD}\hfill \\ \mathrm{LOQ}=10*\mathrm{SD}\hfill \end{array}$$

where: SD – standard deviation of the average blank sample (n = 6).

LOQ of our method was 0.005 mg/L and LOQ was 0.0169 mg/L.

Repeatability and reproducibility

To demonstrate precision of fluoride determination in infusions, both intra-day and inter-day precision were assessed. The inter-day precision (repeatability) and intra-day (reproducibility) were evaluated by measuring six fluoride solutions (n = 6) at one concentration level 1.0 mg/L, under the specific experimental conditions. For repeatability, the analysis was conducted in one day by the same analyst and with the same instrument. The repeatability was expressed as the relative standard deviation (RSD) of six replicates (Table 1).

Table 1.

The inter-day and intra-day precisions

F−concentration	inter-day		intra-day
1.0 mg/L	c* [mg/L] SD [mg/L] RSD recovery n	0.9931 0.0252 2.53% 99.31% 6	c* [mg/L] SD [mg/L] RSD recovery n	0.9857 0.0197 2.00% 98.70% 18

c* – mean fluoride concentration

The intra-day precision was assessed by measuring the same fluoride solutions every day in triplicates for three consecutive days. The reproducibility was expressed by RSD of the average from 18 measurements (Table 1).

Accuracy

Accuracy of the applied analytical method was verified using the CRM material (Certified Reference Material QC3162 Fluoride in Water, Sigma Aldrich) with a certified content of fluoride 0.420 ± 0.0147 mg/L was used. The CRM solutions were prepared in a same way as the samples, i.e. 3 mL of CRM material was mixed with 3 mL of TISAB.

The accuracy was expressed as relative error (%RE). RE indicates the differences between measured value and certified value. The relative error was less than ± 5% in all CRM/TISAB solutions. The results agreed well with the reported certified value of the Certified Reference Material (Table 2).

Table 2.

Results of the CRM measurements

Day	n	Mean [mg/L]	RSD[%]	RE [%]
1	6	0.4171	2.66	2.41
2	3	0.4172	0.91	0.91
3	4	0.4114	0.98	2.04
4	3	0.4195	2.55	1.70
Overall mean		0.4161
Pooled SD		0.0084
Intra SD [%]		2.01
Certified value		0.420 ± 0.0147 [mg/L]

n- number of replicate determinations per day

SD- standard deviation of mean

RSD- relative standard deviation

RE – relative error

Matrix effect

To verify the influence of the interferences and possible matrix effect, the recovery was determined by spiking one randomly selected tea sample (one tea bag, Lipton, black) with a standard solution to get the final concentration 1.0 mg/L. The infusion was prepared in the same way as every other sample (the tea bag was placed in a beaker and poured with 200 mL of boiling water). The 3 µL of standard solution (1000 mg/L) was added to 3 mL of sample and mixed with 3 mL of TISAB prior to measurement. The recovery values were calculated by comparing concentrations obtained from the non-spiked samples F⁻ concentration: 6.759 ± 0.143 mg/L with the spiked ones. The recovery was 98.61%, which confirmed that adding a chelating reagent (TISAB buffer containing CDTA) worked efficiently to avoid the matrix interferences.

Teas and herbs analysis

Based on various manufacture and processing techniques, tea can be categorized into black (fermented, the leaves are almost completely subjected to oxidation by peroxidases and polyphenol oxidases), green (non-fermented, only heated and dried), oolong (partially fermented), red, pu-erh (prolong fermented, but traditionally produced from bigger leaves collected from higher and older trees) and white (un-oxidized, slight fermented, collected only from selected young leaves and petals) [30]. We divided our experimental set into 8 groups: the largest one was the group of black teas (10 products: 6 as tea bags and 4 as leaves). The others were: green teas (5), white teas (3), earl grey (3), pu-erh (3), rooibos (2), yerba mate (2). The investigated herbal samples (5) comprised a separate category. Fluoride was detected in the infusions of all tested teas and herbs and values are given in Table 3.

Our study revealed that fluoride content in the infusions made from teas available on the Polish market varies to a large extent (from 0.718 mg/L to 6.029 mg/L). The highest level was obtained for the infusion of black tea, with the mean fluoride concentration reaching 2.651 ± 1.522 mg/L. Two out of ten tested brands of black tea released a substantial amount of fluoride (4.480 and 6.029 mg/L).

The statistical analysis (one way ANOVA) confirmed the significant differences between groups, p<0.05 (Fig. 2).

Fig. 2.

Fluoride content in infusions of tested products (number of products, n = 33), after 5 min of brewing (p = 0.0136)

Our results obtained for black tea infusions after 5 min brewing were comparable with the data presented by other authors, given in Table 4: 0.6–5.5 [49], 1.15–6.01 [52], 1.47–5.45 [53] and 1.057–6.680 mg/L [46]. The fluoride content at the level of 3 mg/L or lower were also often reported: from 0.56 to 3.06 mg/L [50] 0.57–3.72 mg/L [54], 0.32–2.76 [55], 1.38 [56], 0.53–2.60 [57]. On the other hand, the studies conducted earlier in Poland by Kłódka et al. [58] showed significantly lower fluoride concentrations in black tea from 0.121 to 0.652 mg/L.

Table 4.

Comparison of fluoride concentration in tea infusion [mg/L], based on the literature data

Type of tea	leaves/bags/ stick shape	F concentration in infusion [mg/L]	No. of samples	Tea [g]: water [mL] ratio	Time of brewing [min]	Ref.
Black	bags	1.39–3.06	3	1:100	5	[50]
	stick shape	0.56–1.47	5
	bags	1.40–2.82	11	tea bag:100	5	[54]
	stick shape	0.57–0.94	3
	bags	0.6–5.5	54	tea bag:200	5	[49]
	bags	0.121–0.652	5	0.1:50	6	[58]
	bags	1.15–6.01	25	2:200	10	[52]
	stick	0.95–1.45	6
	bags	2.03–2.76	13	2:100	5	[55]
	leaves	0.32–2.26	13
	bags	2.29–3.28	4		10
	leaves	0.53–2.49	4
	lack of data	0.64–3.55	9	10:400	lack of data	[2]
	bags	1.47–5.45	8	0.5:50	5	[53]
	granular	2.38–3.84	7
	leaves	1.38 ± 0.03	1	1:200	5	[56]
	leaves	2.31 ± 0.05	1		15
	lack of data	0.53–2.60	60	0.2:50	10	[57]
	bags	1.057–6.680	5	1:100	5	[46]
Green	bags	1.40–1.79	3	1:100	5	[50]
	stick shape	0.44–0.92	3
	bags	0.164–0.693	5	0.1:50	6	[58]
	leaves	0.59–1.83	10	2:100	5	[55]
	leaves	0.88–2.14	10		10
	leaves	0.061–0.746	12	2:200	20	[19]
	bag	2.43–6.94	5	0.5:50	5	[53]
	leaves	3.68–5.31	4
	leaves	0.18 ± 0.00	1	1:200	5	[56]
	leaves	0.29 ± 0.01	1		15
	bags	0.2557 ± 0.0542	1	1:100	5	[46]
	leaves	2.177–4.087	4
White	leaves	0.37–0.54	3	2:100	5	[55]
	leaves	0.47–0.59	3		10
	leaves	5.39	1	0.5:50	5	[53]
	leaves	1.07 ± 0.03	1	1:200	5	[56]
	leaves	1.57 ± 0.05	1		15
Earl grey	bags	1.33–3.72	6	tea bag:100	5	[54]
	stick	1.41	1
	bags	2.49–3.55	3	1:100	5	[50]
	leaves	1.22–2.62	2	2:100	5	[55]
	leaves	1.44–3.03	2		10
Pu-erh	bags	0.55 ± 0.02	1	1:100	5	[50]
	stick shape	0.33 ± 0.02	1
	bags	0.43–0.92	3	2:100	5	[55]
	leaves	0.39	1
	bags	0.57–1.04	3		10
	leaves	0.64	1
	leaves	2.87	1	0.5:50	5	[53]
	granular	4.09–4.96	2
	lack of data	0.34–0.690	20	1:50	5	[59]
Rooibos	herbs	0.04–0.06	2	2:100	5	[55]
	bags	0.02–0.03	2
	herbs	0.05–0.08	2		10
	bags	0.04–0.05	2
Yerba mate	leaves	0.03–0.09	2	2:100	5	[55]
	leaves	0.03–0.10	2		10
Chamomile	bags	0.03	1	tea bag:100	5	[54]
	bags	0.03	1		10
	lack of data	1.12	1	1:80	5–30*	[60]
	bag	0.1450 ± 0.0026	1	1:100	5	[46]
Peppermint	bag	0.1707 ± 0.0091	1	1:100	5	[46]
	bags	0.03	1	tea bag:100	5	[54]
	bags	0.03	1		10
Nettle	bags	0.02	1	tea bag:100	5	[54]
	bags	0.03	1		10
Mint (domestic)	lack of data	0.18	1	1:80	5–30*	[60]
Mint (wild)	leaves	12.00	1		5–30*
Mint	leaves	0.002–0.016	15	2:200	20	[19]

*time of brewing according to the manufacturer’s manual for tea preparing

Among the samples of black tea aromatized with bergamot (‘earl grey tea’) the significant amounts of fluoride were extracted, similarly to the black tea samples (Fig. 2). These concentrations were very close to () the values found for non-flavored, “pure” black tea samples (2.359 ± 1.67 mg/L and 2.651 ± 1.52 mg/L, respectively). It seems that an addition of oil from the rind of bergamot oranges (Citrus bergamia) had no influence on fluoride extraction, however the number of samples may be not sufficient to make a viable conclusion. Our highest determined value (4.070 ± 0.054 mg F⁻/L) corresponds with the results found in the literature for earl grey teas: 3.72 mg/L [54] and 3.55 mg/L [50] (Table 4).

The form of tea used for brewing is also meaningful. Our studies showed that significantly higher fluoride contents were measured in black tea infusions made from tea bags than from leaves (Fig. 3.).

Fig. 3.

Fluoride content in infusions of black teas: bags (n = 6, mean = 3.398 mg/L) and leaves (n = 4, mean = 1.529 mg/L, p = 0.0489)

The earlier reports indicated that uptake and accumulation of fluoride in tea leaves increased in time, younger leaves are designated for a production of high-quality tea, while the older ones, richer in fluoride, are used for a cheaper version of the product, such as tea bags, express, bricks or granulated tea [20]. According to Cao et al. [14] brick tea infusions are responsible for adult fluorosis in Tibet irrespective to the fluoride level in drinking water. The similar regularity was also observed in China [61]. Malinowska et al. [55] suggested the use a concentration of fluoride in tea infusion as an indicator/marker of a tea quality. The leaves harvested in early summer are superior in quality to older ones, collected in later seasons, which would be reflected as less fluoride in the infusions. However, more studies should be done to confirm it, particularly in order to compare tea bags and tea leaves, since extraction of fluoride from these two proceed differently. Also the level of fragmentation/fineness can influence the rate of fluoride release from tea material.

An average fluoride concentration in the infusion of the green tea samples was 2.537 ± 1.853 mg/L which was similar to the black tea samples. Only two green tea infusions appeared richer with fluoride, even up to 4.041 and 4.893 mg/L. The average is higher than the results for green tea published by Malinowska et al. [55] ranging from 0.59 to 1.83 mg/L, and by Koblar et al. [50] which were in the range from 0.44 to 1.79 mg/L. Low levels of fluoride in green teas were also reported by other authors [19, 56, 58]. Out of the five samples analyzed in our studies, only two showed a fluoride concentration consisted with a general concept of producing green tea from a bud and two young leaves which are supposed to contain less fluoride than aged leaves. The elevated fluoride concentrations in green teas reported here would require further investigations, as it might be related to a low quality of green teas present on the global market or – again – with the form of tea used here (tea bags, not leaves). However, there are available data about fluoride concentration similar to our results: 0.2557–4.087 in Japanese green teas [46] or even higher: 2.43–6.94 mg/L in Chinese and American teas [53] (Table 4).

More delicate and younger tea leaves are used for production of white tea; this has been reflected in the 2–3 fold-decreased fluoride concentration; the mean value for the white tea samples was 1.913 ± 0.687 mg/L. In the literature data on the white tea, a low F⁻ level is reported, such as 0.37–0.54 [55] and 1.07 mg/L [56], but there is also one, particularly high content of 5.39 mg/L [53], which is closer to the one in a black tea infusion.

In the case of pu-erh (so-called red tea), our results were similar to white tea varying between 1.186 and 2.634 mg/L, with the mean value 1.996 ± 0.739 mg/L (Table 3). Other authors presented a comparable fluoride content in infusion from pu-erh leaves, i.e. 2.87 mg/L, or even higher (from pu-erh in granular form) i.e. 4.09–4.96 mg/L [53], but there were also lower values (Table 4) [50, 55, 59].

Herbal infusions, in comparison to traditional black or green teas, contained almost negligible amounts of fluoride, with a mean value 0.0572 mg/L for all studied herbs, ranging from 0.043 ± 0.002 mg/L (chamomile) to 0.077 ± 0.001 mg/L (mint). The values reported by other authors about chamomile, nettle and mint at the level 0.016–0.03 mg/L (Table 4) [19, 54] are consistent with our results (Table 3). In the other studies, the values were not always comparable, from 0.145 mg/L for chamomile and 0.170 mg/L for peppermint [46], 0.18 mg/L for domestic mint, 1.12 mg/L for chamomile and extremely high, 12.00 mg/L for wild mint [60].

Additionally, it is worth noting that there is absolutely no available data on herbs as common as yarrow (Achillea millefolium) and purges (Cistus incanus).

Comparing rooibos infusions (from 0.038 to 0.090 mg/L) and yerba mate (from 0.041 to 0.045 mg/L) we observed a similarly low level of fluoride. These results are consistent with the literature data [55], although the publications related to the fluoride content in infusion of rooibos and yerba mate are rather sparse. It appears, that only the ‘original tea plant’ (Camellia sinensis L.) has truly selective possibility to accumulate fluoride. The two above-mentioned popular beverages, despite being called “tea”, are not produced from the tea plant (Camellia sinensis L.), but from other species (rooibos – Aspalathus linearis, yerba mate – Ilex paraguariensis).

Influence of brewing time

Brewing time of 5 min is most often recommended by tea manufacturers and this time span was selected as the first variant in our studies, although tea consumers usually do not strictly respect the time of brewing, which may result in elution of additional quantities of fluoride. The second brewing time used in our experiments was 15 min. After the prolonged time of brewing the same tendency of fluoride concentrations in the tested groups of tea was observed, and the differences between various types of tea were statistically significant (Fig. 4).

Fig. 4.

Fluoride content in infusions of the tested products (number of products n = 33), after 15 min of brewing (p = 0.0091)

Prolonging the brewing time from five to fifteen minutes increased the fluoride extraction from black tea by 12 to even 36% (Fig. 5.). A similar observation was reported by other researches (Table 4) [55, 56]. In the case of green and white tea, the fluoride concentration was also elevated, but not that much, i.e. by 0.3 to 15.8% and by 5.7–9.6%, respectively. The highest growth was observed at the lowest concentration level, meaning in herbal, rooibos and yerba mate infusions.

Fig. 5.

Percentage increase in fluoride content after additional 10 min of brewing

The longer brewing time of herbs also led to increasing a concentration of fluoride, although it remained (as before) very low, comparing to the original tea products. The level of fluoride in the herbal infusions remained lower than in the tea extracts by about two orders of magnitude (Table 3), despite the time of brewing.

In our studies, after 5 min of brewing, the highest fluoride concentration was observed in the infusions of mint leaves, and it was increasing by about 11% after additional 10 min. The other herbs reacted nearly the same way (nettle and chamomile by 27% and 30%, respectively), except for purges, where almost a double increase of fluoride concentration in herbal was determined. The infusions of yarrow kept steady concentrations of fluoride, both during 5 and 15 min of brewing. Emekli-Alturfan and co-workers [54] confirmed the increase of fluoride concentration only in case of nettle, from 0.02 to 0.03 mg/L. The other studied herbs (chamomile and peppermint) did not show such a trend and the fluoride level was stable. It must be emphasized though, that the fluoride contents in the herb infusions are so extremely low, that increasing the brewing time will not make the beverages dangerous for human health.

According to WHO, the optimum fluoride intake for humans should range from 2 to 4 mg/day [3]. The level recommended by US EPA is similar, from 2.5 to 4 mg F⁻/day for children and adults, respectively [62]. This amount of fluoride can be covered by about 2–3 cups (ca. 400–600 mL) daily of a popular black tea with F⁻ concentration about 6 mg/L, even after 15-minute brewing, but without taking into account other sources of fluoride (e.g. toothpaste etc.).

Conclusions

Infusions made from plants (different types of tea, tea-like products and herbs), which are capable to accumulate fluoride, may constitute an important source of this element supplied for a human body and it can be considered as an substantial part of a healthy diet. Infusions of herbal products, as well as rooibos and yerba mate contain almost negligible concentrations of fluoride, contrary to the products prepared out of the tea plant (Camellia sinensis L.), which is proved to be a “specific fluoride accumulator”. All the tested tea infusions (of black, green, white and red tea) contain substantial amounts of fluoride. It was related to the type/form of tea or the way of fermentation. Our results showed the general rank with respect to the fluoride concentration in an infusion:

• black tea > green tea > earl grey > pu-erh > white tea>>>rooibos, yerba mate, herbal products

Moreover, the longer the brewing time, the higher the concentration is observed in the infusion, particularly of black tea, rising from 12 to 36%. It must be emphasized that the amount of fluoride in the analyzed infusions of teas and herbs was not high enough to cause a risk of fluorosis, even if left to brew up to 15 min. Nevertheless, a fluoride intake from teas and herbs consumed regularly should be taken into account in view of the potential health effects, both positive and negative, especially for children, and pregnant women who often choose plant infusions as healthier replacement of coffee. Above all, in the total fluoride balance, it is necessary to take into account the fluoride content in water used for brewing, which may be significant and variable depending on the country and other sources of fluoride.

Although the confirmed levels were not alarmingly high, due to constantly changing environment and the methods of cultivation of tea and other herbal plants used for the preparation of daily consumed infusions, the awareness of harm obligate to a constant monitoring of the fluoride content. Similarly to the publications related to tea consumption in other societies (considered as “heavy tea drinkers”) it is emergent to verify if this process can be considered as a health risk or not. Our experiments revealed that even with a highly fluoride-enriched tea infusion or longer than necessary tea brewing, fluoride exposure is negligible for Polish society. Moreover, in our screening we added the most popular herbs, which are prepared such as tea infusion and willingly consumed. The fluoride level appeared insignificant for health risk assessment, which gives an opportunity in the future to look for tea replacement (particularly now, when the world-wide pandemic can compromise trade exchange across the world).

Author contributions

The first author named is lead and corresponding author. We describe contributions to the paper using the taxonomy provided above. Conceptualization: A.S. and E.A.S.; Methodology: A.S.; Validation: A.S. and A.K.; Formal Analysis: A.S. and A.K.; Investigation: A.S. and A.K.; Resources: A.S.; Writing – Original Draft: A.S. and A.K.; Writing – Review & Editing: A.S., A.K. and E.A.S.

Declarations

Disclaimer

The reported names of tea and herbal brands were used only to presenting the values and cannot be used to introduce customers to buy a given product. None of measured values pose threat to public health. The study does not provide any health-based arguments for buying specific brand of product.

Conflict of interest

The authors declare that they have no conflict of interest.