Fluoride Action Network

Fluoride, Water Hardness, and Endemic Goitre

The Lancet | May 27, 1972

May 27, 1972; Pages 1135-1138


East Wing, Guy’s Hospital, London S.E.1


The prevalence of goitre in 17 Himalayan villages has been estimated. Water-samples from each village were taken, and levels of iodine, fluoride, and hardness determined. In 13 villages wide variations in goitre prevalence were not attributable to differences in iodine intake, which remained constant within a narrow range. Instead, variations in goitre prevalence were found to correlate closely with the fluoride content (p=0-74; P<0-01) and with the hardness (p=0.77; P<0-01) of the water in each village. The effects of fluoride and water hardness seem to be independent.


Goitre prevalence in a number of endemic areas has been shown to depend on factors other than iodine deficiency (1-3).

The role of water hardness in endemic goitre in man and experimental goitre in animals is well established (1). That fluoride may be another important factor has long been suspected. Goitre has been produced in experimental animals by feeding fluoride salts (5-7), and fluoride exerts an antithyroid effect in man when used in the treatment of thyrotoxicosis (8,9).

However, epidemiological evidence of an association between fluoride and endemic goitre is contradictory. While areas in which the water contains an unusually high level of fluoride, or in which dental fluorosis is pronounced, have been shown to coincide with the regional distribution of endemic goitre (10-13), the absence of endemic goitre in other areas where the fluoride levels are equally high has been noted (14-16).

It has been suggested that a high intake of iodine may offset the goitrogenic effect of fluoride, and this might explain the absence of goitre in some high fluoride areas. Iodine levels were consequently measured during surveys (4,17-20), in which goitre prevalence in centres of relatively high and low fluoride intake was compared. Unfortunately, iodine intake was generally high, and varied widely from one centre to another, so no true estimate could be made of the independent effect of fluoride. Several workers (4,18,20) felt that the possible goitrogenic effect of fluoride coule only be usefully studied in an area of low iodine intake and of high goitre prevalence.

Our study was performed in such an area, and provides data upon the relation between fluoride and endemic goitre under conditions of constant low iodine intake. In addition, further evidence is provided in support of the association between endemic gotire and water hardness.


Area of Survey

The survey was conducted in Chintang panchayat, Nepal, in conjunction with a B.C.G. inoculation programme. The panchayat covers an area of approximately 25 square miles of hill-country of 500-6000 ft., with a population of 10,000 people.

The population is uniform with respect to race, occupations, and eating habits. The diet is based upon rice, pulse, vegetables, and spices. It is low in iodine and lacks the common vegetable goitrogens. Iodised salt is not available locally.

17 villages were visited and 736 people examined. 4 villages had relatively high water-iodine levels and were excluded. Our study is based upon data from the remaining 13 (648 people) which had a water-iodine concentration of 0-001 p.p.m. or less.

Sample Selection and Goitre Assessment

The prevalence of goitre in each village was estimated on a sample representing about a third of the population. The samples consisted of all children attending for B.C.G. inoculation plus all accompanying adults. The same selection procedure was followed in each village.

Goitres were graded into three sizes by inspection, according to the method of Perez et al. (21), the age and sex of each subject being noted.

Water Sampling and Analysis

Each of the villages in the survey depended upon one source of water, in most cases a soil well. A sample of water was taken from each source into a 50 ml. High density polyethylene bottle, which had been washed in distilled water. The bottles did not contaminate or absorb ions from test solutions.

Estimations of fluoride, calcium, magnesium, and iodine were performed in London, after completion of the goitre survey, in the laboratory of the Government Chemist. Fluoride was determined by the specific-ion electrode (22); calcium and magnesium by atomic absorption spectrophotometry; and iodine by Dubravcic’s modification of the ceric-ion-reduction technique (23). Water hardness was derived from the calcium and magnesium levels, and expressed as the concentration of a pure solution of calcium carbonate having the same degree of hardness as the sample.


The prevalence of clearly visible goitre – grades two and three of Perez’ classification (21) was calculated for each village. Grades two and three include only those goitres which are easily discernible without extension of the neck.

The age and sex distribution of goitre in the 13 villages is given in table I. There is a pronounced rise in prevalence with age, but the sexes are affected almost equally, which is characteristic of a hyperendemic area (1).

Though a constant very low iodine intake was common to the people of all the villages, goitre prevalence varied widely from village to village. This variation correlated closely with corresponding variations in both water hardness and fluoride content (P <0-01) (table II). Since water hardness and fluoride levels are themselves intercorrelated, it is not possible on these data to decide whether the association between fluoride and goitre is independent of that between hardness and goitre. A more complex analysis of the data involving the use of partial correlation coefficients would not be justified, as assumptions would need to be made about the linearity of association and the normal distribution of variables. However, some indication that the effect of fluoride is independent of that of hardness is provided when the data are grouped as in the accompanying figure. It can be seen that the lowest goitre prevalence occurs in soft-water/low-fluoride villages, whilst goitre prevalence is higher when the water either is harder or contains more fluoride. The highest prevalence of all is in villages with both hard-water and high-fluoride levels.

Unlike McCarrison (l), we found no striking inverse relation between altitude and goitre prevalence. Nor was there an obvious association between the type of water source and goitre. All villages were supplied by shallow soil wells except F, H, and M (streams) and J (river).


Water Hardness

The occurrence of endemic goitre in hard-water areas has often been noted (4). Experimental work in rats supports the idea (24) that calcium salts exacerbate an underlying state of iodine deficiency. Where iodine is not plentiful, the hardness of water may be more important in determining goitre prevalence than the absolute level of iodine: in the U.K., hard waters, though containing a greater concentration of iodine, are associated with a higher goitre prevalence than soft (4). Our results show that, with constant low-iodine levels, a very close correlation may be demonstrated between the degree of hardness and goitre prevalence.

Although the goitrogenic effect of calcium salts has been elucidated by experiment (4), the action of magnesium ions on thyroid function has been little investigated. In our data magnesium levels are more closely correlated with goitre prevalence than calcium levels, but this does not necessarily mean that magnesium is a more potent goitrogen. Indeed, the correlation may reflect only the fact that the two elements tend to occur together in the rocks and soil of the region of the survey.


Suspicions of an aetiological connection between fluoride and goitre, raised by the observation of a high incidence of goitre in areas of endemic fluorosis (10-13), have not been substantiated by epidemiological studies (17-20) conducted in relatively non-goitrous regions.

However, the possibility that fluoride may influence the prevalence of goitre in an area where goitre is endemic has not previously been investigated. Our results show a close quantitative association between fluoride and goitre under these circumstances.

This association is unlikely to be due to chance, but it may be indirect, and not the result of fluoride exerting a goitrogenic effect. That the association is probably causal is suggested by the fact that fluoride produces goitre when fed to animals (1-7). However, the doses of fluoride used in animal experiments have been large compared to the doses normally consumed by man (25). Further experimental work is needed, using lower doses of fluoride, and iodine-deficient partially goitrous animals, to justify the conclusion that fluoride is normally goitrogenic under circumstances of iodine deficiency.

Experimental work should also be directed towards discovering the site of action of fluoride in thyroid function. Many workers have suggested that it affects iodine metabolism, but there is little conclusive evidence of such a mechanism (26,27).

Possibly the influence of fluoride is upon the aminoacid precursors of thyroxine-tyrosine and its metabolites-rather than upon iodine. Increased urinary loss of tyrosine is known to occur in men living in a high-fluoride area (28) and in monkeys receiving low daily doses of fluoride (29): tyrosine deficiency has been shown to cause thyroid hypofunction in rats (31). Moreover, the coexistence of fluorosis and goitre has been noted especially in poor rural populations consuming diets deficient in protein, and the other features of fluorosis-tooth and bone changes-are enhanced by malnutrition (26, 31).

Proportion with goitre as a fraction:
Village By age (yr.) By sex Total Prevalence (%)
0-5 6-12 13-18 19+ M F
A 0/11 1/13 0/6 1/8 0/21 2/17 2/38 5
B 0/20 4/19 0/4 2/5 1/22 5/26 6/48 12
C 0/9 1/10 0/5 3/7 1/19 3/12 4/31 13
D 1/11 2/14 0/8 4/7 1/19 6/21 7/40 17
E 1/13 2/22 7/12 0/3 2/19 8/31 10/50 20
F 1/12 2/12 2/4 4/7 3/15 6/20 9/35 26
G 0/8 2/13 0/3 8/12 5/19 5/17 10/36 28
H 2/16 0/13 3/8 14/16 7/21 12/32 19/53 36
I 0/19 10/28 11/18 19/20 11/35 29/50 40/85 47
J 2/11 13/29 11/23 19/26 26/52 19/37 45/89 51
K 2/12 6/14 3/4 12/13 13/23 10/20 23/43 53
L 1/9 9/18 8/9 11/12 16/26 13/22 29/48 60
M 3/10 14/20 8/10 11/12 22/26 14/26 36/52 69
Totals 13/161 66/225 53/114 108/148 108/317 132/331 240/648
(8%) (29%) (46%) 73% (34%) (40%) (37%)

Water Concentration in parts per million (p.p.m) of: Hardness as CaC03
Village Goiter Prevalence (%) I F Ca Mg (p.p.m)
A 5 0.001 <0.1 6 2.0 24
B 12 <0.001 0.13 7 1.5 24
C 13 <0.001 0.13 3 0.5 10
D 17 0.001 0.12 6 1.5 21
E 20 <0.001 0.19 69 15.5 235
F 26 <0.001 0.24 108 28.5 385
G 28 <0.001 0.22 8 3.5 33
H 36 <0.001 0.28 73 5.0 203
I 47 <0.001 0.21 12 6.0 54
J 51 0.001 0.19 148 65 630
K 53 <0.001 0.36 14 17 102
L 60 0.001 0.23 121 39 458
M 69 <0.001 0.23 145 77 670
Spearman’s p = 0.74 0.78 0.83 0.77
P < 0.01 <0.01 <0.01 <0.01



We thank Dr. S. J. Patterson and Mr. N. G. Bunton, of the Government Chemist’s laboratory, for performing the water analysis; His Majesty’s Government and the people of Nepal, and the Britain-Nepal Medical Trust, for help in Nepal; and Dr. Margaret Crawford and Mr. David Clayton for help with the statistics. The work was supported by a Searle scholarship of the British Medical Students’ Trust.

Requests for reprints should be addressed to T. K. D.


1 . McCarrison, R. The Simple Goitres. London, 1928.
2. Young, M., Crabtree, M. G., Mason, E. M. Spec. Rep. Ser. med. Res. Coun. 1936, no. 217.
3. PediatricsSpringfield, 1967, 40, 935.
4. Murray, M. M., Ryle, J. A., Simpson, B. W., Wilson, D. C. Med. Res. CounMemo. 1948, no. 18.
5. Maumene, M. E. C. r. hebd. Seanc. Acad. Sci. Paris, 1854, 39, 538.
6. Roholm, K. Fluorine Intoxication; p. 89. Copenhagen, 1937.
7. Hodge, H. C., Smith, F. A. in Fluorine Chemistry (edited by J. H. Simons); vol. iv, p. 106. London, 1965.
8. Goldemberg, L. Presse med. 1930, 102, 1751.
9. May, W. Klin. Wschr. 1935, 14, 790.
10. von Fellenberg, T. Mitt. Lebensmitt. Hyg. 1938, 29, 276.
11. Wilson, D. C. Lancet, 1941, i, 211.
12. von Hoffmann-Axthelm, W. Dt. zahnirztl. Z. 1953, 8, 757.
13. Steyn, D. G. Leech, Johannesb. 1964, 34, 114.
14. Demole, V., von Fellenberg, T., Held, A. J., Schmid, H. Bull. schweiz. Akad. med. Wiss. 1951, 7, 440.
15, Leone, N. C., Shimkin, M. B., Arnold, F. R., Stevenson, C. L., Zimmermann, E. R., Geiser, P. B., Lieberman, J. E. Publ. Hlth Rep. Wash. 1954, 69, 925.
16. Frada, G., Mentesana, G., Guajana, U. Minerva medRoma, 1969, 60, 545.
17. Wespi, H. J. Praxis, 1954, 28, 616.
18. Berry, W. T. C., Whittles, J. H. Mon. Bull. Minist. Hlth, 1963, 22, 50.
19. Gedalia, L, Brand, N. Archs int. Pharmacodyn. Thir. 1963, 142, 312.
20. Siddiqui, A. H. Y. Endocr. 1960, 20, 101.
21. Perez, C., Scrimshaw, N. S., Munoz, J. A. Monograph Ser. W.H.O. 1960, 44, 369.
22. Patterson, S. J., Bunton, N. G., Crosby, N. T. Wat. Treat. Examination, 1969, 18, 182.
23. Dubravcic, M. Analyst, Land. 1955, 80, 295.
24. Thompson, J, Archs Path. 1933, 16, 211.
25. Demole, V. Bull. schweiz. Akad. med. Wiss. 1954, 10, 292.
26. Ericsson, Y. Br. dent. J. 1967, 123, 276.
27. Report of the Royal Commission into Fluoridation; p. 153. Hobart, 1968.
28. Singh, A., Vazirani, S. J., jolly, S. S., Bansal, B. C. Postgrad. med. J. 1962, 38, 150.
29. Pandit, C. G., Rao, D. N. Indian J. med. Res. 1940, 28, 559.
30. Simon, G., Szilcs, J., Gyetvai, G., Kecskem6ti, V. Acta physiol. hung. 1962, 21, 335.
31. Murray, M. M., Wilson, D, C. Br. dent. J. 1948, 84, 97.