Limitations There are various limitations to our study. We only studied AN people who volunteered from one region of southwest Alaska in Bristol Bay. Therefore, the results from this regional study may such information not be representative of the general population of AN people. Those who enrolled were principally Yupik, and there are many Alaska Native people in other areas of Alaska, who come from different tribes and ethnic backgrounds. Yupik people primary inhabit the west coast of Alaska and make up over 30% of the Alaska Native population (Census 2010, http://www2.census.gov/geo/maps/special/AIANWall2010/AIAN_AK_2010.pdf). They are the largest Alaska Native tribal grouping, either alone or in combination with other races (34,000). Yupik also had the greatest number of people who identified with one tribal grouping and no other race (29,000; Census 2010, http://2010.

census.gov/news/releases/operations/cb12-cn06.html). Even though the number of participants was relatively small, our paper provides new findings about these unique rural dwellers. There are no other published data on age of first exposure, risk perception, types of tobacco used including dual use, or nicotine dependence scores. Iqmik users were difficult to recruit because there are fewer numbers of AN who use this product in the recruitment villages; however, Yupik people in other regions have iqmik prevalence rates as high as 22% among adults and 18.3% of Alaska Native women reported prenatal ST or iqmik use during 2008 (AK DHSS, 2009; ANEC, ANTHC, 2009).

Another limitation is the small sample sizes in certain categories of tobacco users, so statistical comparisons may not be informative due to limited power. Furthermore, participants enrolled in this study may not be generalizable to different subgroups of tobacco users and only reflect the characteristics of subjects interested in participating in a study. Dependence scores were assessed using scales developed for heavier smokers. It may be that the lower scores reflect a general low level of tobacco use, but not of low levels of dependence. In summary, multiple product use and early onset did not appear correlated with heavier tobacco use or high levels of addiction (using available dependence scales) in this population. As suggested, patterns of disparities in tobacco product use, exposure, and associated disease are complex and involve interactions among a range of factors (Fagan et al.

, 2007). Biological samples collected will provide further understanding of differences in exposure to nicotine and toxicants. The results from these analyses will further our understanding of product content and the impact of patterns of consumption and individual differences in rates of metabolism on levels of exposure. Dacomitinib Our study will also help guide future studies and programs of tobacco cessation interventions among AN people.