Potassium fluoride

Epidemiological data

Administrative data

Endpoint:

epidemiological data

Type of information:

other: human data

Adequacy of study:

supporting study

Reliability:

other: not rated acc. to Klimisch

Rationale for reliability incl. deficiencies:

other: Well-documented publication

Data source

Reference

Reference Type:

publication

Title:

Fluoride in drinking water and risk of hip fracture in the UK: a case control study

Author:

Hillier, S et al.

Year:

2000

Bibliographic source:

The Lancet 355: 265 - 269.

Materials and methods

Study type:

other: population-based case-control study

Endpoint addressed:

repeated dose toxicity: oral

Test guideline

Qualifier:

no guideline followed

Principles of method if other than guideline:

A case-control study of men and women aged 50 years and older from the English county of Cleveland, and compared patients with hip fracture with community controls was conducted. Current addresses were ascertained for all participants; for those who agreed to an interview and who passed a mental test, more detailed information was obtained about lifetime residential history and exposure to other known and suspected risk factors for hip fracture. Exposures to fluoride in water were estimated from the residential histories and from information provided by water suppliers. Analysis was by logistic regression.

GLP compliance:

not specified

Test material

Test material information

Constituent 1

Reference substance name:

fluoride in drinking water

IUPAC Name:

fluoride in drinking water

Method

Type of population:

general

Ethical approval:

confirmed, but no further information available

Details on study design:

SETTING: County of Cleveland in northeast England
The Hartlepool area of Cleveland had always been supplied with water naturally high in fluoride (>1 ppm), whereas the rest of the county received water with a low fluoride concentration.

STUDY POPULATION
The study population consisted of residents of Cleveland.
- Selection criteria: virtually all cases of hip fracture in the county are treated at three hospitals—Hartlepool General Hospital, North Tees General Hospital (Stockton), and Middlesbrough General Hospital. The researchers identified all members of the study population who were admitted to these hospitals during a 17-month period with newly diagnosed fractures of the femoral neck that were through or above the lesser trochanter and not caused by cancer. Patients who agreed to participate and achieved a score of six or higher on a Hodkinson abbreviated mental test (Quereshi & Hodkinson, 1974)* were allowed to participate in the study.
- Total population (Total no. of persons in cohort from which the subjects were drawn): 914 subjects
- Total number of subjects participating in study: 514 subjects
- Sex/age: men and women; 50 years and older

COMPARISON POPULATION
- Type: Control or reference group
The control group was randomly selected from a list of all members of the study population registered with National Health Service (NHS) general practitioners, and controls were frequency matched to the cases for age (in 5-year bands) and sex. The control group was matched only by age and sex, and not by area of residence within Cleveland. Those people who agreed to participate and who scored 6 or more on the Hodkinson abbreviated mental test were interviewed. The same questionnaire was used for both cases and controls.
- Total population (Total no. of persons in cohort from which the subjects were drawn): 1196 subjects
- Total number of subjects participating in study: 527 subjects

METHOD OF DATA COLLECTION
- Type: Questionnaire
- Details: the questionnaire asked about demographic variables, height, weight, lifetime residential history, usual physical activities, age at menopause, alcohol consumption, smoking history, recent medication, and dietary sources of calcium and fluoride. From the reported height and weight the body-mass index was calculated. The measures of physical activity (eg, walking speed, time spent walking outside, and time spent gardening) were highly
intercorrelated, and were combined to give a summary physical activity score that was grouped into three categories. The questions on dietary calcium were used to classify intake as low, medium, or high.

To test whether estimates of exposure to fluoride in drinking water were linked to the amount of fluoride in femoral bone, the researchers collected the excised femoral heads from a subset of 105 cases (25 in Hartlepool, 80 in Stockton and Middlesbrough) who were treated by arthroplasty. The samples were analysed without knowledge of the fluoride exposure of the patient, according to a previously described method (Ishiguro et al, 1993)*. A 1 cm slice of trabecular bone was cut from the femoral head, and if the exercised bone included part of the femoral neck, a piece of cortical bone was also obtained. Slivers of about 50 mg were dissolved in 1 mol/L perchloric acid, and the fluoride concentration was measured with a fluoride electrode. The amount of fluoride in the bone was given as mol/g dry weight and was linked to the patient’s estimated average lifetime exposure to fluoride in drinking water.

*References:
- Quereshi KN, Hodkinson HN. Evaluation of a ten question mental test in the institutionalised elderly. Age Ageing 1974; 3: 152–57.
- Ishiguro K, Nakagaki H, Tsuhoi S, et al. Distribution of fluoride in cortical bone of human rib. Calcif Tissue Int 1993; 52: 278–82.

Exposure assessment:

estimated

Details on exposure:

TYPE OF EXPOSURE MEASUREMENT:
The residential histories of cases and controls were reviewed by a person who was not aware of their case/control status, and information provided by water suppliers was used to assign a water fluoride concentration for each year of the person’s life. When this information was acquired for at least 90% of years, it was collated to obtain estimates of the average fluoride concentrations to which the person had been exposed throughout life, up to age 20 years, and in the 20 years before entry to the study. In some areas, including Cleveland, measurements of fluoride in water had been done from before 1960. In other areas, measurement did not begin until the early 1980s. However, concentrations of fluoride before the start of monitoring could be estimated with reasonable confidence when sources of supply had not changed.

Statistical methods:

The associations of hip fracture with exposure to fluoride and other risk factors were analysed by logistic regression with adjustment throughout for age and sex.

Results and discussion

Results:

EXPOSURE
Data on lifetime exposure to fluoride in drinking water were at least 90% complete for 460 cases and 423 controls. The average lifetime concentrations that cases and controls had been exposed to ranged from 0.15 ppm to 1.79 ppm. Most values were either less than 0.2 ppm (76% of participants) or 0.9 ppm and higher (18%), reflecting the large difference between Hartlepool and elsewhere in Cleveland and the low rate of migration between different parts of the county. Current residence in Hartlepool was a good indicator of lifetime exposure to fluoride in water.
Please also refer to table 1 in the field "Any other information on results incl. tables" below.

FINDINGS
-The following findings were made when looking at the associations of hip fracture with lifetime exposure to fluoride in drinking water, and also with exposures early in life and more recently: risk estimates were all close to unity, and were virtually unchanged by adjustment for potential confounders. The adjusted odds ratio associated with an average lifetime exposure to fluoride in drinking water of 0.9 ppm or higher was 1.0 (95% CI 0.7–1.5). There was no indication that fluoride was associated with a higher risk of hip fracture than normal in people with low intakes of dietary calcium.
Please also refer to table 2 in the field "Any other information on results incl. tables" below.

The authors also calculated the risk of fracture, adjusted for age and sex, by current place of residence, and in this analysis they included all cases and controls, whether or not they were interviewed. By comparison with those living elsewhere, the odds ratio for residents of Hartlepool was 1·0 (0·8–1·3).

The main dietary source of fluoride in the UK is tea (tea leaves contain high concentrations of fluoride). After adjustment for the potential confounding factors (body-mass index, physical-activity score, age at menopause, current alcohol consumption, smoking, current treatment with oral corticosteriods and dietary intake of calcium), no significant association was seen between hip fracture and higher reported tea consumption (odds ratio 1·3 [95% CI 0·9–2·0]).

Fluoride content was analysed in 79 samples of trabecular bone and 30 samples of cortical bone from patients for whom data on lifetime exposure to fluoride in drinking water were at least 90% complete. For trabecular bone there was no clear association between exposure to fluoride and the amount of fluoride in the bone. In all four samples of cortical bone from patients exposed to high concentrations of fluoride in water, however, there was an increased amount of fluoride in the bone.

Confounding factors:

- analysis allowed for possible confounding by variousknown ans suspected casues of fracture.
- Relation of hip fracture to variables that might confound an association with fluoride:
Strong associations were recorded with low body-mass index (p for trend <0.001) and low degrees of physical activity (p for trend <0.001). There was an increased risk in patients who were being treated with oral corticosteroids, although this increase was not significant. By contrast, no clear associations were seen with age at menopause, alcohol consumption, smoking, or dietary calcium intake (p for trend 0.5).

Strengths and weaknesses:

Weakness
- limitation of low participation rate in interviews

Strength:
- exposures during extended periods should be more relevant to risk than exposures assessed onyl at a single point in time

Any other information on results incl. tables

Table 1: Lifetime exposure of study participants (cases and controls combined) to fluoride in drinking water according to current residence

Average concentration of fluoride in drinking water	Current residence
	Hartlepool	Elsewhere
<0-9 ppm	20 (12%)	706 (99%)
≥0-9 ppm	150 (88%)	7 (1%)
Total	170 (100%)	713 (100%)

The mean fluoride concentrations in the two exposure categories (for all areas of current residence combined) were 0.2 ppm and 1.5 ppm.

Table 2: Associations of hip fracture with exposure to fluoride in drinking water

Period of exposure and fluoride concentration (ppm)	Number of cases	Number of controls	Odds ratio (95% Cl)
			Adjusted for age and sex	Adjusted for all potential confounders†
Lifetime exposure
<0.9	380	346	1.0*	1.0*
≥0.9	80	77	0.9 (0.7-1.3)	1.0 (0.7-1.5)
Exposure to age 20 years
<0.9	397	393	1.0*	1.0*
≥0.9	77	67	1.1 (0.8-1.6)	1.1 (0.7-1.7)
Exposure in previous 20 years
<0.9	402	391	1.0*	1.0*
≥0.9	85	87	0.9 (0.7-1.3)	0.9 (0.6-1.4)

*Reference category. †Includes age, sex, and potentially confounding variable (body-mass index, physical-activity score, age at menopause, current alcohol consumption, smoking, current treatment with oral corticosteriods and dietary intake of calcium).

Applicant's summary and conclusion

Conclusions:

The authors did not find evidence of any increase in the risk of hip fracture from fluoride in drinking water at concentrations of about 1 ppm.