Perchlorate exposure has been associated with decreased thyroxine and increased thyroid stimulating hormone in women with lower iodine intakes in the U.S. population 
. Further analyses find that low iodine intake coupled with concurrent exposure to multiple iodide uptake inhibitors (e.g. perchlorate, thiocyanate and nitrate) may decrease thyroid function 
. Turkey has a history of low iodine intake as well as potentially significant exposure to perchlorate and other iodide uptake inhibitors 
. Therefore we designed this pilot study to investigate the prevalence of low iodine intake coupled with concurrent exposure to perchlorate, thiocyanate and nitrate. We found that the median urinary perchlorate concentration (6.4 µg/L) was more than twice as high as the median perchlorate concentration found in U.S. women (2.9 µg/L) 
. Similarly, the median perchlorate dose across all Turkish sites (0.13 µg/kg/day) was 2.6 times higher than the median perchlorate dose found in U.S. women (0.051 µg/kg/day) 
. Median perchlorate dose was below the U.S. EPA reference dose (0.7 µg/kg/day), but nine study participants had perchlorate doses higher than the U.S. EPA reference dose (see Figure S2 in File S1
. Further study is needed to explore the potential impact of these perchlorate exposures.
The sources of perchlorate exposure in the study population are not known. Perchlorate enters the environment from both natural and anthropogenic sources and is stable in arid soils and water, leading to environmental persistence 
. Food and forage crops can uptake perchlorate from soil and irrigation water, leading to human exposure from consuming the food crops or from consuming milk produced by cattle fed perchlorate-contaminated forage crops 
. Thus, foods (leafy vegetables, milk products, and fruits) and drinking water may be significant contributors to perchlorate exposure in Turkey as well. Across the three cities studied, Isparta had lower perchlorate concentrations and doses compared with Kayseri (p<0.05; ). Lower perchlorate exposure in Isparta could result from differences in locally grown food or local water disinfection practices 
. Additional data are needed to characterize perchlorate exposure sources in Turkey.
The recommended iodine intake for women of reproductive age is 150 µg/day 
. The range of iodine excretion measured in 24 hr urine indicated that few of the study population consumed adequate levels of iodine (see Figure S1 in File S1
). Populations are considered to have adequate iodine intake if the median urinary iodine levels are between 100−199 µg/L according to the WHO (World Health Organization) 
. Our results (median urinary iodine
67.1 µg/L) agree with other studies that find that the Turkish population is moderately iodine deficient 
, . We found lower median levels of urinary iodine (67 µg/L) compared with a recent study by Erdogan et al (2007) that measured median iodine levels (107 µg/L) in morning urine samples of school-age children from 24 cities and from 7 regions in Turkey 
. In the one city (Istanbul) that was sampled in both studies, Erdogan et al (2007) found twice the level of urinary iodine (154 µg/L compared to 77.5 µg/L). This difference in urinary iodine levels is attributable to the age of the study participants: children tend to have much higher urinary iodine levels compared with adults 
. In fact, urinary iodine data from NHANES consistently finds that women of reproductive age have about half the urinary iodine levels compared with children 
. In contrast to NHANES data indicating adequate iodine intake in the US population, we found inadequate iodine intake (), suggesting ongoing iodine deficiency in all three cities studied.
The public health strategy to reduce iodine deficiency is salt iodization; therefore we expected higher iodine levels in urine collected from people who consume iodized salt. Urinary iodine levels were marginally higher in women using iodized salt (67.9 µg/L) compared with women not using iodized salt (47.8 µg/L), although this difference was not significant once we controlled for age, BMI and study site. According to the Turkey Demographic and Health Survey (2008) 15% of the households did not have iodized salt; furthermore, the availability of iodized salt differed by residence type and region 
. In urban areas, only 10% of the household salt tested was not iodized, whereas this value goes up to 30% in rural areas. Despite the fact that 91% of the study participants reported using iodized salt in our study, the observed low levels of iodine intake indicate that additional efforts are needed to protect the Turkish population from iodine deficiency.
Istanbul participants were younger and of lower BMI than study participants from the other two locations. These demographic differences might affect the results. Previous reports indicate that people with higher BMI tend to excrete higher levels of perchlorate and other food-related anions 
. Similarly, older U.S. adults tend to excrete more perchlorate than do younger U.S. adults, although the reason for this observation is not clear 
. We controlled for differences in age and BMI between the three cities by using multivariate models. After adjusting for differences in age and BMI, urinary nitrate levels were lower in Isparta (21850 µg/L) compared with Kayseri (47450 µg/L, ). The higher nitrate levels observed in Kayseri may result from higher levels of nitrate in local food and drinking water. Indeed the City of Kayseri Municipal Water and Sewer facility has reported nitrate levels as high as 49.25 mg/L 
, raising concerns about potential health effect 
. Further work is needed to characterize nitrate exposure sources and health effects in Turkey.
Multivariate analysis found that smokers had significantly higher cyanide exposure compared with non-smokers (p<0.0001). The effect of smoking on the urinary thiocyanate levels is illustrated in . Urinary thiocyanate levels increased with increasing cigarettes smoked per day (CPD), with heavy smokers (>10 CPD) having higher urinary thiocyanate levels (2410 µg/L) compared with light smokers (≤10 CPD, 1110 µg/L), who had higher urinary thiocyanate levels compared with non-smokers (268 µg/L). These higher thiocyanate levels are indicative of higher exposure to cyanide gas from tobacco smoke. Median thiocyanate levels in all three groups of Turkish women were lower than median levels in US women in the National Health and Nutrition Examination Survey (NHANES) 2001−2002 (1260 µg/L) 
, perhaps because Turkish women smoke fewer cigarettes compared with US women.
The scatter plot matrix illustrates correlations among analytes (). Perchlorate, nitrate and iodine were more tightly correlated with each other than with thiocyanate, likely because of differences in exposure sources. Perchlorate, nitrate and iodine exposures are likely from the same sources (for example, dairy products tend to contain significant levels of both perchlorate and iodine, and leafy vegetables tend to contain significant levels of both perchlorate and nitrate). Conversely, tobacco smoke was the primary source of urinary thiocyanate as a metabolite of the cyanide in the tobacco smoke. We further explored second hand smoke exposure at home or at work as a potential source of thiocyanate, but did not find secondhand smoke categorization to be significantly related to increased urinary thiocyanate levels. Detailed distributions of tobacco smoke exposure results are shown in Figure S4a, Figure S4b and Figure S4c in File S1
This pilot study provides novel data indicating that study participants had low iodine intake and high intake of some iodide uptake inhibitors compared with reference populations. However, the study also is weak in that it draws these conclusions based on a relatively small number of participants (N
255) and possibly biased selection between study sites. Thus, our findings need to be confirmed in larger groups of participants, especially in pregnant and lactating women. While the study does use rigorous 24-hr urine collection, multiple 24-hr samples would have resulted in more precise exposure estimates. Additionally, the study would have been strengthened by full assessment of current thyroid function of study participants.
Individuals with lower ratios of iodine to iodide uptake inhibitors may be more prone to iodide uptake inhibition, with perchlorate, nitrate and thiocyanate possibly out-competing iodide for transport into the thyroid 
. Chronically low levels of iodine relative to iodide uptake inhibitors could lead to decreased thyroid hormone production. Although our data only provide a 24-hr snap shot of the relative levels of iodide and iodide uptake inhibitors, it identifies lower levels of iodine and higher levels of perchlorate compared with U.S. reference data. Thus, iodide uptake may more likely be inhibited in this population compared to the U.S. population. For these reasons we aim to perform further studies to determine the sources of these contaminants, and to relate exposures to thyroid hormone levels.