Fluoride and health : epidemiological studies of fluoride exposure and hip fracture, myocardial infarction and osteosarcoma

Sammanfattning: The aim of this thesis was to investigate the association between drinking water fluoride exposure and risk of hip fracture, myocardial infarction, and osteosarcoma. Swedish nation- wide population-based registers have been used throughout the thesis. The risk of hip fracture was addressed in a population-based cohort of 452,824 eligible individuals with an estimated exposure to the same drinking water source from birth upon start of follow-up, (i.e. living in their municipality of birth). Information on residence from parish records was used to address a community water supplier for each individual in the cohort. The drinking water fluoride levels in our cohort ranged between <0.1 and 2.7 mg/L. We linked the Total Population Register, the National Patient Register, and the Swedish Death Register. We studied the association between drinking water fluoride level and risk of hip fracture. We found no association between fluoride exposure level and risk of hip fracture (compared to the very low exposure group, adjusted Hazard Ratio (aHR) for the low exposure group was 0.97; 95% CI = 0.94-0.99, aHR for the medium exposure group was 0.97; 95% CI = 0.94-1.00, and aHR for the highest exposure group was 0.98; 95% CI 0.93- 1.04). Nor did we find an association between fluoride level and the risk of osteoporotic (low- trauma) hip fracture. Stratified analyses suggested that fluoride exposure in individuals younger than 80 years of age was associated with a decreased risk for hip fracture. However, no clear exposure-response effect was observed. We cannot rule out that unmeasured confounding may have influenced the observed results. The risk of myocardial infarction was addressed in a population-based cohort of 455,619 eligible individuals with an estimated exposure to the same drinking water source from birth upon start of follow-up, (i.e. living in their municipality of birth). The fluoride exposure assessments and the retrieval of register data were performed in a similar manner as in study I (described above). The drinking water fluoride levels in our cohort ranged between <0.1 and 2.7 mg/L. We studied the association between drinking water fluoride level and risk of myocardial infarction. We found no association between fluoride exposure level and risk of myocardial infarction (compared to the very low exposure group, aHR for the low exposure group was 1.00; 95% CI 0.99-1.02, aHR for the medium exposure group was 1.02; 95% CI 0.99-1.04, and aHR for the highest exposure group was 1.01; 95% CI 0.98-1.04). Additional analyses was performed, looking at fatal and non-fatal myocardial infarction. No association was found. We found some evidence of a positive association between fluoride exposure level and risk of myocardial infarction in the northern part of Sweden (compared to the very low exposure group, aHR for the low exposure group was 1.04; 95% CI 1.00-1.07, aHR for the medium exposure group was 1.12; 95% CI 1.07-1.16, and aHR for the highest exposure group was 1.09; 95% CI 1.02-1.17). This association may in part be explained by unmeasured confounding. Moreover, only a small effect size is detected, and in combination with the large sample size and total number of events, we consider that this finding is unlikely to be of significance. The risk of osteosarcoma was addressed in a population-based case-control study consisting of 363 eligible osteosarcoma cases identified in the Swedish Cancer Register, and 1,815 control subjects randomly selected from the Total Population Register, and matched to cases (5:1) on birth year and sex. A control had to be alive and without an osteosarcoma diagnosis at the time of selection. All domestic movements and migrations (dates and locations) were collected for all cases and controls, from birth upon index date (date for cancer diagnosis). Information on residence was used to address a community water supplier for each individual in the cohort. The drinking water fluoride levels in our cohort ranged between 0.03 and 2.75 mg/L (ppm). We linked the Swedish Cancer Register, the National Patient Register, the Total Population Register, and The Register of Population and Population Changes. We studied the association between accumulated drinking water fluoride exposure and risk of osteosarcoma. Different exposure-times before osteosarcoma diagnosis were studied, as well as different exposure-ages, to evaluated potential important time-windows for exposure. We found no association between drinking water fluoride exposure and risk of osteosarcoma (aOR 0.99; 95% CI 0.67-1.16). However, this study has a few important limitations. We lack information on potential confounders such as additional socioeconomic variables and height at diagnosis. Moreover, misclassification of exposure when we only assess fluoride exposure from drinking water, other dietary source for fluoride (food, beverages, and dental hygiene products) might account for an important part of the total fluoride load. And additionally, non-differential misclassification of exposure (i.e. individuals categorized as exposed could in fact be unexposed and vice versa). We cannot rule out that the above mentioned methodological limitations may have influenced the results. In conclusion, fluoride exposure from municipal drinking water does not appear to be associated with increased risk for hip fracture, myocardial infarction, or osteosarcoma in Sweden. This is in agreement with most findings from previous studies assessing similar drinking water fluoride exposure levels. The results of this thesis add to the body of evidence that ingestion of lower fluoride concentrations (<4 mg/L) is not associated with increased risk of adverse health effects.