- Research article
- Open Access
- Open Peer Review
Big toenail and hair samples as biomarkers for fluoride exposure – a pilot study
BMC Oral Health volume 19, Article number: 82 (2019)
Biomarkers can aid in detecting and preventing clinical disease through the recognition of change in biological samples. The objective of this case-control study was to further the knowledge on the use of big toenail and hair samples as biomarkers for fluoride exposure.
A total of 48 participants from an endemic (IC) and a non-endemic (SC) fluorosis region were included. Big toenail and hair samples were collected from each participant and the fluoride concentration was determined. The results of 42 participants were reported: 21 participants (11 males and 10 females, mean age 15.66 + 2.61 years) from IC and 21 participants (11 males and 10 females, mean age 15.06 + 0.79 years) from SC.
The mean fluoride concentration of big toenail (2.34 ± 0.26 mg/kg) and hair (0.24 ± 0.04 mg/kg) in the endemic region was significantly higher than the mean fluoride concentration of big toenail (0.98 ± 0.08 mg/kg) and hair (0.14 ± 0.02 mg/kg) in the non-endemic region (p < 0.001 and p = 0.004, respectively). The Receiver Operating Characteristic (ROC) analysis showed that the Area Under the Curve (AUC) value was 0.889 for big toenail (p < 0.001) and 0.762 for hair (p = 0.004) samples. The fluoride assay for big toenails exhibits greater observed accuracy than does the fluoride assay for hair.
Nail and hair samples can serve as biomarkers to detect biological fluoride exposure according to the data of this pilot study. Nevertheless, hair is less sensitive and specific as a biomarker when AUC values of big toenail and hair samples were compared.
Biomarkers, i.e. biological markers, can aid in detecting and preventing clinical disease through the recognition of change in biological systems or samples . Specifically, they have been defined as “cellular, biochemical, or molecular alterations which are measurable in biological media such as human tissues, cells, or fluids and are indicative of exposure to environmental chemicals” , such as fluoride.
Bone [2, 3], dentin [3, 4], plasma, saliva and urine [5,6,7] as well as nails [3, 8,9,10,11,12] and hair [3, 13, 14] are biomarkers that were investigated to assess fluoride exposure. It has been pointed out that among the mineralized tissues, bone is the main site for fluoride accumulation, thereby making bone a good choice as a fluoride biomarker . However, the difficulty and the invasiveness of bone sample collection has been underlined. Thus, the collection of dentin, from extracted third molars, has been suggested as more appropriate when compared to bone sample collection .
Less invasive methods, using body fluid samples, such as plasma, saliva and urine also have been considered in the literature to analyze body fluoride concentration [5,6,7]. Nevertheless, these body fluids are affected by a number of variables, such as fluoride intake within the last few hours. Consequently, these body fluids present short-term, i.e. ‘snapshot’, information only [3, 7].
Nail samples, on the other hand, can be obtained non-invasively. They can be easily transported and stored for long periods of time without degradation [3, 8,9,10,11]. Nails, particularly from the big toe (hallux), providing enough mass for fluoride analysis as well as their faster growth rate when compared to the other toenails, have been recommended as suitable biomarkers for fluoride intake [3, 10]. Also, toenails have been reported to be less prone to external contaminants when compared to fingernails [3, 10, 11].
The rationale for the use of hair as a suitable biomarker for fluoride is the same as that for toenail and fingernail clippings . The endogenous trace element composition of hair and nails is believed to reflect the metabolic milieu during their formation . The concentration of fluoride in hair as well as nail samples represents the average level of fluoride intake and plasma concentration over an extended period of time [3, 8, 9, 13].
It was pointed out that some individuals may not accept hair sample collection, particularly individuals with long hair, since hair sample collection has to be carried out as close as possible to the scalp . In these individuals big toenail clipping samples might serve as an alternative.
Different methodologies have been used for the determination of fluoride in biological materials, such as nail and hair samples. Yet, the most popular analytical method for the extraction of fluoride from biological samples has been reported to be the hexamethyldisiloxane (HMDS)-facilitated diffusion method. This technique has been described as simple and fast [8, 9, 16].
The aim of this case-control study with a 1:1 allocation ratio was to test the null hypothesis that big toenail and hair samples of subjects, living in endemic and non-endemic fluorosis regions, cannot serve as biomarkers for fluoride exposure.
Sample size calculation was performed using Pocock’s formula for two means . With 20 participants per group, the trial has 80% power to detect a clinically meaningful difference of 1.25 mg per kilogram (mg/kg) of fluoride in big toenail (hallux) clippings between the two regions at the 5% significance level. In order to overcome drop out and exclusion from the study, 24 participants were included in each arm of the study.
Ethical approval was granted by the Medical Faculty Ethics Committee of Ondokuz Mayis University, Turkey (No. 2008/143). The study sample included forty-eight participants who came from two regions, i.e. an endemic and non-endemic fluorosis region. The trial was undertaken with the understanding and written consent of each participant/guardian. All participants stated that they were lifelong residents in their respective areas. No changes in the address of residency (< 3 months) particularly before the commencement of this study and no history of systemic or topical fluoride supplements existed. Twenty-four participants (12 males and 12 females, mean age 15.42 + 2.50 years) were from Isparta city (IC), in the southwestern part of Turkey, with a high fluoride concentration in the public water supply (≥2 ppm (ppm)) [18, 19]. The participants from IC had Thylstrup and Fejerskov Fluorosis Indices (TFI) ranging from 2 to 5 . The remaining 24 participants (12 males and 12 females, mean age 15.15 + 0.96 years) came from Samsun city (SC), on the north coast of Turkey, which has a low fluoride concentration (≤0.05 ppm) in the public water supply . The Thylstrup and Fejerskov Fluorosis Indices (TFI) were 0 for the participants form SC . Criteria for selection of big toenail clipping and hair samples are given in Table 1.
Sampling and analyses
At the beginning of this study, the participants were instructed to let their big toenails and hair grow prior to sample collection for four weeks and six weeks, respectively.
On the scheduled day, the participants/parents visited their respective dental school to enable the investigator (SET) to collect big toenail and hair samples. Occipital hair is the only source recommended for the analysis for both male/female subjects. High-grade stainless steel scissors were utilized to cut the hair. The hair tuft sample was collected at a distance of 0.5 centimeters (cm) from the scalp and approximately 3 cm in length. The weight required for the hair specimen was 50–100 milligrams (mg). Hair samples were stored in labelled polyethylene bags, in a dry place, at room temperature. Again, the investigator (SET) clipped the big toenail’s free end and cut from the right side to the left side using each participant’s nail clipper. The clipped nails were stored separately in labelled plastic boxes per participant. The nail and hair samples were initially forwarded to the University of Sydney, Australia. Subsequently, the samples were sent to Bauru School of Dentistry, University of São Paulo, Brazil for analyses.
Fluoride concentration in nail and hair samples was determined after overnight HMDS-facilitated diffusion, applying the Taves method  as modified by Whitford  using a fluoride ion-specific electrode (Orion Research, Cambridge, Mass., USA, model 9409) and a miniature calomel reference electrode (Accumet, No. 13–620–79), both coupled to a potentiometer (Orion Research, model EA 940). All readings were made in duplicate.
Statistical analyses were performed with SPSS 23.0 for windows. The mean repeatability of the readings, based on duplicate samples, was 96%. Data were presented as mean ± standard error (SE) and median (Semi Interquartile Range – SIQR). The Shapiro–Wilk test was used to analyze the normal distribution assumption of the quantitative outcomes. Mann-Whitney test was used to compare the fluoride concentration in big toenail as well as hair samples between endemic and non-endemic regions. The receiver operating characteristic (ROC) curve was used to illustrate and evaluate the performance of big toenail and hair samples as biomarkers in case of fluoride exposure. The area under the ROC curve (AUC) was evaluated as the measure of a diagnostic test’s discriminatory power. Confidence intervals can be computed for AUC. In this article, sensitivity and specificity values were evaluated. A p value less than 0.05 was considered as statistically significant.
The fluoride concentration in 6 participants could not be detected (3 from each city) due to technical difficulties, i.e. the fluoride concentration was not within the detection limit of the electrode since the amount of the sample collected was insufficient. Thus, 21 residents from IC (11 males and 10 females, mean age 15.66 + 2.61 years) and 21 residents from SC (11 males and 10 females, mean age 15.06 + 0.79 years) remained for the final analyses.
Overall, the mean fluoride concentration collected from nail clippings of the residents of IC was 2.34 + 0.26 mg/kg, while it was 0.98 + 0.08 mg/kg for those from SC (Table 2). Hair sample analysis of the participants from IC displayed a fluoride concentration of 0.24 + 0.04 mg/kg, while it was 0.14 + 0.04 mg/kg for those from SC (Table 2). There was a significant difference in the concentration of fluoride in nail clipping (p < 0.001; Table 3; Fig. 1) and hair (p = 0.004; Table 3; Fig. 2) samples between the two cities, higher in participants from IC than those from SC.
The results of ROC analysis are given in Table 4 and Fig. 3. The relative positions of the plots indicate the relative accuracies of the tests. A plot lying above and to the left of another plot indicates greater observed accuracy. In Fig. 3, the fluoride assay for big toenails exhibits greater observed accuracy than does the fluoride assay for hair.
The results of ROC analysis showed that the area under the curve (AUC) of big toenail samples was 0.889 (p < 0.001). The optimal sensitivity and specificity were 0.810 and 0.857, respectively (Table 4, Fig. 3). The AUC for hair samples was 0.762 (p = 0.004). The optimal sensitivity and specificity were 0.762 and 0.762 respectively (Table 4, Fig. 3).
The relationship between fluoride concentrations in big toenail (hallux) clippings, hair and the level of fluoride in the water of an endemic (IC) and non-endemic (SC) fluorosis region was assessed in this study.
The results of the present study, namely a higher and significant fluoride concentration in nail clippings and hair collected from IC participants when compared to SC participants, are similar to the findings of other studies [11,12,13,14]. These findings may be attributed to the fact that the regular level of fluoride in IC (≥2 ppm) drinking water was higher than that in SC (≤0.05 ppm) [18, 19]. Therefore, systemic fluoride circulation in the residents of IC is considerably higher, which in turn raised the fluoride uptake by nails and hair.
Absorbed fluoride deposits in the growing toenail by either continuous incorporation or secondary concentration . One might argue that the detected level of fluoride in the nail clippings and hair collected from IC had a wide standard deviation indicating a significant individual discrepancy. This trend might be related to factors such as variation in the amount of consumed water [11,12,13,14], diet [14, 22], tea consumption  and external contamination among participants  as well as interpersonal variation in the amount of absorbed, circulated, metabolized and deposited fluoride .
The area under the ROC curve showed that nail and hair samples can serve as a biomarker to detect biological fluoride exposure according to the data of this pilot study. AUC is an effective way to summarize the overall diagnostic accuracy of a test . In general, an AUC value of 0.7 to 0.8 is considered acceptable; whereas, an AUC value of 0.8 to 0.9 is considered excellent . Thus, nail and hair samples have a reasonable discriminating ability to diagnose fluoride exposure. Nevertheless, hair is less sensitive and specific as a biomarker when AUC values of big toenail and hair samples are compared.
It is worth noting that the fluoride concentration for hair was lower than that for toenails. This outcome might be due to the fact that hair is characterized by a cyclic growth rate with different stages , whereas nails grow continuously and do not have a growth cycle analogous to that of hair . In contrast, the fluoride concentration in toenails was higher than hair fluoride concentration and almost equal to the associated water. The growth rate of toenails  is substantially lower when compared to the growth rate of hair . This lower growth rate of toenails might allow a more significant accumulation of fluoride. Furthermore, incorporation of fluoride through the nail bed, not only through the matrix (growth end), might contribute to the total fluoride concentration in toenails [3, 31].
The null hypothesis was rejected, i.e. nail and hair samples can serve as biomarkers to detect biological fluoride exposure according to the data of this pilot study. Nail and hair samples have a reasonable discriminating ability to diagnose fluoride exposure from the water supply from an endemic and non-endemic fluorosis region. Nevertheless, hair is less sensitive and specific as a biomarker when AUC values of big toenail and hair samples were compared. This area merits further research with a larger sample size.
Area Under the Curve
Centimeter per month
Milligram per kilogram
Millimeter per day
Parts per million
Receiver Operating Characteristics
Thylstrup and Fejerskov Fluorosis Index
Hulka BS, Wilcosky T. Biological markers in epidemiologic research. Arch Environ Health. 1988;43:83–9.
Turner CH, Boivin G, Meunier PJ. A mathematical model for fluoride uptake by the skeleton. Calcif Tissue Int. 1993;52:130–8.
Pessan JP, Buzalaf MAR. Historical and recent biological markers of exposure to fluoride. Monogr Oral Sci. 2011;22:52–65.
Vieira AP, Hancock R, Limeback H, Maia R, Grynpas MD. Is fluoride concentration in dentin and enamel a good indicator of dental fluorosis? J Dent Res. 2004;83:76–80.
Ekstrand J. A micromethod for the determination of fluoride in blood plasma and saliva. Calcif Tissue Res. 1977;23:225–8.
Usuda K, Kono K, Dote T, Nishiura K, Miyata K, Nishiura H, et al. Urinary biomarkers monitoring for experimental fluoride nephrotoxicity. Arch Toxicol. 1998;72:104–9.
Rugg-Gunn AJ, Villa AE, Buzalaf MAR. Contemporary biological markers of exposure to fluoride. Monogr Oral Sci. 2011;22:37–51.
Whitford GM, Sampaio FC, Arneberg P, von der Fehr FR. Fingernail fluoride: a method for monitoring fluoride exposure. Caries Res. 1999;33:462–7.
Whitford GM. Monitoring fluoride exposure with fingernail clippings. Schweiz Monatsschr Zahnmed. 2005;115:685–9.
Buzalaf MAR, Pessan JP, Alves KMRP. Influence of growth rate and length on fluoride detection in human nails. Caries Res. 2006;40:231–8.
Fukushima R, Rigolizzo DS, Maia LP, Sampaio FC, Lauris JRP, Buzalaf MAR. Environmental and individual factors associated with nail fluoride concentration. Caries Res. 2009;43:147–54.
Sankhala SS, Harshwal R, Paliwal P, Agarwal A. Toe nails as a biomarker of chronic fluoride exposure secondary to high water fluoride content in areas with endemic fluorosis. Fluoride. 2014;47:235–40.
Mandinic Z, Curcic M, Antonijevic B, Carevic M, Mandic J, Djukic-Cosic D, et al. Fluoride in drinking water and dental fluorosis. Sci Total Environ. 2010;408:3507–12.
Parimi N, Viswanath V, Kashyap B, Patil PU. Hair as biomarker of fluoride exposure in a fluoride endemic area and a low fluoridated area. Int J Trichology. 2013;5:148–50.
Hopps HC. The biologic bases for using hair and nail for analyses of trace elements. Sci Total Environ. 1977;7:71–89.
Whitford GM. In: Myers HM, editor. Absorption and plasma concentrations of fluoride. Basel: Karger: The metabolism and toxicity of fluoride; 1996. p. 26–7.
Pocock SJ. The size of a clinical trial. In: Pocock SJ, editor. Clinical trials: a practical approach. Chichester: John Wiley & Sons; 1983. p. 125–9.
Davraz A, Sener E, Sener S. Temporal variations of fluoride concentration in Isparta public water system and health impact assessment (SW-Turkey). Environ Geol. 2008;56:159–70.
Karadeniz EI, Gonzales C, Turk T, Isci D, Sahin-Saglam AM, Alkis H, et al. Effect of fluoride on root resorption following heavy and light orthodontic force application for 4 weeks and 12 weeks of retention. Angle Orthod. 2013;83:418–24.
Thylstrup A, Fejerskov O. Clinical appearance of dental fluorosis in permanent teeth in relation to histological changes. Community Dent Oral Epidemiol. 1978;6:315–28.
Taves DR. Separation of fluoride by rapid diffusion using hexamethyldisiloxane. Talanta. 1968;15:969–74.
Chowdhury C, Khijmatgar S, Kumari DP, Chowdhury A, Grootveld M, Hedge C, et al. Fluoride in fish flesh, fish bone and regular diet in south-coastal area of Karnataka state of India. Indian J Dent Res. 2018;29:414–7.
Das S, de Oliveira LM, da Silva E, Liu Y, Ma LQ. Fluoride concentrations in traditional and herbal teas: health risk assessment. Environ Pollut. 2017;231:779–84.
Czarnowski W, Krechniak J. Fluoride in the urine, hair, and nails of phosphate fertiliser workers. Br J Ind Med. 1990;47:349–51.
Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010;5:1315–6.
Hosmer DW, Lemeshow S. Assessing the fit of the model. In: Cressie NAC, Fisher NI, Johnstone IM, Kadane JB, Scott DW, Silverman BW, et al., editors. Applied logistic regression. New York: John Wiley & Sons; 2000. p. 160–4.
Alonso L, Fuchs E. The hair cycle. J Cell Sci. 2006;119(Pt3):391–3.
Baumgartner MR. Nails: an adequate alternative matrix in forensic toxicology for drug analysis? Bioanalysis. 2014;6:2189–91.
Yaemsiri S, Hou N, Slining MM, He K. Growth rate of human fingernails and toenails in healthy American young adults. J Eur Acad Dermatol Venereol. 2010;24:420–3.
Myers RJ, Hamilton JB. Regeneration and rate of growth of hairs in man. Ann N Y Acad Sci. 1951;53:562–8.
Corrêa Rodrigues MH, de Magalhães Bastos JR, Rabelo Buzalaf MA. Fingernails and toenails as biomarkers of subchronic exposure to fluoride from dentifrice in 2- to 3-year-old children. Caries Res. 2004;38:109–14.
The corresponding author (SET) is particularly grateful to the Department of Orthodontics, Suleymen Demirel University, Isparta for providing the necessary compartment for sample collections.
This research was partially funded by the Australian Society of Orthodontists Foundation for Research and Education (ASOFRE). Specifically, the transfer of the samples was funded by the ASOFRE. ASOFRE had no role in the study design, data collection, data analysis, data interpretation, or writing of the manuscript.
Availability of data and materials
The data of this research is available from Selma Elekdag-Turk (corresponding author)/ Associate Professor/ Department of Orthodontics, Faculty of Dentistry/ Ondokuz Mayis University, Samsun, Turkey as well as from M. Ali Darendeliler / Professor and Chair/ Discipline of Orthodontics/ Faculty of Dentistry/ University of Sydney, Australia upon request.
Ethics approval and consent to participate
Ethical approval was obtained from the Medical Faculty Ethics Committee of Ondokuz Mayis University, Samsun, Turkey (No. 2008/143). This research was undertaken with the understanding and written consent of each participant/guardian. Furthermore, the participants of this research were informed that the results will be published.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.