Environmental epidemiology of type 1 diabetes mellitus in Prince Edward Island

The main objective of this thesis was to examine the relationship between type 1 diabetes mellitus (T1D) and dietary nitrate intake from both drinking water and food sources in Prince Edward Island (PEI), Canada at the ecological level, and at the individual level in a case-control study. Historical data on ground water nitrate concentrations and land use patterns were utilized to examine temporal and spatial assumptions made in these two T1D studies. Ground water nitrate concentrations were assessed temporally, and the association between ground water nitrate and local land use was assessed spatially, comparing areal aggregation methods. The relationship between average nitrate concentration in ground water and T1D incidence at the watershed level was assessed, taking into account the population-at-risk and average household income. A case-control study compared drinking water chemistry, food frequency and average dietary component (e.g. nutrients) consumption between patients diagnosed with T1D during a four year period, and their age and sex matched controls, with an emphasis on nitrate concentrations and its derivatives. Some environmental and genetic-based factors were evaluated and controlled for where appropriate.Average monthly values in nitrate-nitrogen concentration for 54 wells across PEI over a three year period ranged from 3.80 mg L-1 in April to 4.08 mg L-1 in January, and annual values ranged from 2.12 mg L-1 in 1987 to 2.73 mg L-1 in 1983 for 167 wells over a 16-year period, with monthly measurements significantly differing over time. Local land use had a major influence on average nitrate concentrations: agricultural areas, particularly where row-crops grew, had higher nitrate concentrations than both residential and 'pristine' areas, and this was somewhat dependent on season. Specifically, potato, grain, and hay land coverage were significantly and positively related to average ground water nitrate concentrations (compared to 'pristine' areas), regardless of method of data aggregation. Blueberry land coverage was negatively associated with ground water nitrate concentrations using two of the three aggregation methods. Watersheds were the preferred method of nitrate aggregation at the area level when compared to freeform polygons and 500 m buffer zones around each well, due to the method of creation (hydrological boundaries), being large enough to accurately determine average nitrate concentration, and being able to explain the most amount of nitrate concentration variation. The results from these historical data indicate that the assumptions made in the T1D incidence and risk studies were valid: there were minimal temporal differences but substantial spatial differences in ground water nitrate concentrations.The incidence of T1D at the watershed level was not associated with ground water nitrate concentrations either before or after adjusting for average household income (a proxy for socioeconomic status), but a weak trend of higher nitrate concentrations being positively associated with T1D risk was present, especially after controlling for average household income.A case-control study comprising of 57 T1D cases and 105 controls, matched by age at diagnosis and sex, concluded that the risk of T1D was increased when regular soft drinks or eggs were consumed at 'least once per week', compared to 'less than once per week'. The intake of dietary nitrate was marginally significant and positively associated with T1D (p = 0.13). An increased intake of total carbohydrates and caffeine were also associated with an increased risk of T1D, whereas the increased intake of vitamin A, vitamin B12, folate, and zinc were associated with a decreased risk. A family history of T1D and having five or more infections during the first two years of life significantly increased the risk of T1D, whereas residential remoteness was associated with a reduced risk of T1D. Father's education was also related to the risk of T1D, but the direction of the relationship was not clear.In summary, ground water nitrate concentrations in PEI were influenced by local land use more than temporal factors. Nitrate from food had a marginally significant dose-response relationship with T1D risk at the individual level, and T1D incidence was positively but not significantly associated with nitrate concentrations in ground water at the area level. Other environmental and dietary factors had a greater influence on T1D risk at the individual level.Source: Dissertation Abstracts International, Volume: 68-01, Section: B, page: 0186