Matern Child Health J DOI 10.1007/s10995-014-1475-2
Fish Consumption and Hair Mercury Levels in Women of Childbearing Age, Martin County, Florida Anil Nair • Melissa Jordan • Sharon Watkins • Robert Washam • Chris DuClos • Serena Jones • Jason Palcic • Marek Pawlowicz • Carina Blackmore
Ó Springer Science+Business Media New York 2014
Abstract The health effects of mercury in humans are mostly on the developing nervous system. Pregnant women and women who are breastfeeding must be targeted in order to decrease mercury exposure to the populations at highest risk—infants, unborn fetuses, and young children. This purpose of this study is to understand the demographics of fish-consumption patterns among women of childbearing age (including pregnant women) in Martin County, Florida, and to analyze the associations of mercury levels in participants’ hair with socio-demographic variables in order to better design prevention messages and campaigns. Mercury concentrations in hair samples of 408 women ages 18–49 were assessed. Data on demographic factors, pregnancy status, fish consumption, and awareness of fish advisories were collected during personal interviews. Data were analyzed using descriptive statistics and multivariate logistic regression. The geometric and arithmetic means of hair mercury concentration were 0.371 and 0.676 lg/g of hair. One-fourth of the respondents had a concentration C1 lg/g of hair. Consuming a higher number of fish meals per month, consumption of commercially purchased or locally caught fish higher in mercury, White race and income C$75,000 were positively associated with A. Nair M. Jordan (&) S. Watkins C. DuClos C. Blackmore Bureau of Epidemiology, Florida Department of Health, 4052 Bald Cypress Way, Bin A-08, Tallahassee, FL 32399, USA e-mail: [email protected] R. Washam S. Jones Martin County Health Department, Florida Department of Health, 3441 SE Willoughby Boulevard, Stuart, FL 34994, USA J. Palcic M. Pawlowicz Bureau of Public Health Laboratories, Florida Department of Health, 1217 Pearl Street, Jacksonville, FL 32202, USA
the likelihood of having higher hair mercury levels. This study confirms the existence of a higher overall mean hair mercury level and a higher percentage of women with C1 lg/g hair mercury level than those reported at the national level and in other regional studies. This suggests the need for region-specific fish consumption advisories to minimize mercury exposure in humans. Keywords Fish consumption Hair Mercury Women of childbearing age
Introduction Mercury is a metal existing naturally in the earth’s crust. Through a combination of natural and anthropogenic processes, mercury gets deposited in water and becomes methyl mercury through biotransformation by aquatic microorganisms [1, 2]. Fish consumption is the most important pathway of mercury exposure in humans. While all fish have methyl mercury in their tissues, predatory fish that are long-lived tend to have higher body burdens of methyl mercury [1, 3]. The most important toxic effect of methyl mercury is on the development of the nervous system [4]. Methyl mercury can pass through the placental barrier and affect the nervous system of developing fetuses [5]. Developmental neurotoxicity has also been observed following prenatal exposure of children to lower levels of methyl mercury from maternal fish consumption. Studies have reported the adverse developmental effects of high levels of maternal methyl mercury in newborns [6–8]. Using national blood mercury prevalence data from the Centers for Disease Control and Prevention, it was estimated that between 316,588 (7.8 %) and 637,233 (15.8 %) of infants born in the U.S. annually would have cord blood
123
Matern Child Health J
mercury levels [5.8 lg/L [9], a level associated with a decline in IQ. In adults, total blood mercury concentrations reflect methyl mercury exposures over several months, but may be influenced by recent consumption of fish or shellfish meals [10]. Hair sample analysis is a common method used to detect exposure to mercury over several months to a year prior to sampling [11]. Hair concentration is reported to correlate well with dietary methyl mercury intake and hair sample collection is simple and non-invasive [12]. Hair mercury has been used as a biomarker for determining methyl mercury exposure among fish-eating populations [12–15]. Considering the potential adverse effects of methyl mercury, the United States Environmental Protection Agency (USEPA) and the Food and Drug Administration (FDA) have jointly provided fish consumption advisories. Likewise, some states have also provided advisories to limit fish consumption by women of childbearing age. The Florida Department of Health (FDOH) recommends pregnant and nursing women and women who may become pregnant not eat certain fish species, such as: Chilean sea bass, golden snapper, jack, king mackerel, marlin, orange roughy, sea lamprey, shark, Spanish mackerel, sword fish, tile fish, tuna and walleye [16]. According to the World Health Organization [17], dependence on fish as a protein source is higher in coastal than in inland areas. Variation in the mercury content in fish from different regions of Florida has also been reported [18]. The objectives of the present study were to understand the demographic and fish-consumption patterns of women of childbearing age in Martin County, Florida, and to analyze the associations of mercury levels in participants’ hair with socio-demographic variables in order to better design prevention messages and campaigns.
Table 1 Participant demographic characteristics compared with the 2010 U.S. Census Bureau’s reported demographic characteristics for Martin County, Florida [n (%)]
Methods
a
Total sample (n = 408a), women living in Martin County, FL Ages 18–49
Martin County, FL total population = 146,393 U.S. Census Bureau (2010)
Age (years) 18–24
51 (12.5)
10,687 (22.5)
25–34
122 (29.9)
10,979 (23.1)
35–49
235 (57.6)
25,912 (54.5)
345 (84.6) 22 (5.4)
128,510 (88.8) 8,775 (6.0)
Race White Black Asian
7 (1.7)
904 (0.6)
Native American
5 (1.2)
1,765 (1.2)
28 (6.9)
6,439 (4.4)
\25,000
67 (16.4)
12,018 (20.0)
25,000–34,999
39 (9.6)
9,795 (16.3)
35,000–49,999
53 (13.0)
8,953 (14.9)
Other Income ($)b
50,000–74,999
71 (17.4)
9,254 (15.4)
136 (33.3)
20,010 (33.3)
12 (2.9)
14,159 (12.9)
High school/ GED
46 (11.3)
27,989 (25.5)
Some college/ technical school
119 (29.2)
25,136 (22.9)
229 (56.1)
42,478 (38.7)
C75,000 Educationc Elementary/ middle school
College graduate/post college
Due to missing values, not all subgroups total to 408
b
Human Subjects Review
Census data for income are reported by number of households at each level, not individuals
c
The study protocol was submitted to the FDOH Ethics and Human Research Protection Program in April 2009 to determine if this mercury biomonitoring activity involved research. It was determined that this study was designed to establish baseline mercury levels for purpose of public health surveillance and is not primarily an activity designed to contribute to generalizable knowledge, so review by the FDOH Institutional Review Board was not required.
Census data on educational attainment are only available for those ages 25 and older
had resided in Martin County, Florida for at least 1 year. A recruitment announcement for women was posted in newspapers, physician offices, health clubs, health department clinics and websites, and advertised on local public television stations. Although not a random sampling, the final demographic characteristics of respondents were not markedly different from the overall county population [19] (Table 1).
Sample Selection Data and Hair Sample Collection We performed a cross-sectional study in a medium sized coastal county in Florida. The data for this study were collected in 2010 from 408 volunteer women ages 18–49, who
123
Upon verification of the eligibility criteria, the selected volunteers were asked to sign an informed consent form
Matern Child Health J
outlining the details of the project. Data on a variety of factors, including: fish consumption habits, knowledge of fish consumption advisories, pregnancy status, demographic and socio-economic information and mercury exposure history were collected from the selected respondents using a pre-tested interview form. The interview questions had previously been used in two related studies in Florida. At the end of the interview, the respondents were asked to provide approximately 50 mg (*100 strands) of their hair sample cut close to the scalp from the occipital region of the head using stainless steel surgical scissors. Hair samples were analyzed for concentration of total mercury by the FDOH Bureau of Public Health Laboratories in Jacksonville by combustion-gold amalgamation-atomic absorption spectroscopy (C-GA-AAS) with the MA-2000 mercury analyzer (Nippon Instruments Corporation, Tokyo, Japan). The calibration linear dynamic range was modified (2- to 60-ng of mercury) to achieve lower detection limits than previously reported [20]. Statistical Analysis All analyses were performed using SAS software (version 9.2, SAS Institute, Carey, NC) and a p value of B0.05 was considered significant in all analyses. Descriptive statistics were used to present the frequency distribution of the number of participants according to age group, race, educational attainment, annual household income, possession of a fishing license, fish consumption patterns, awareness about limiting fish consumption, awareness about local fish advisories, consumption of commercially purchased or locally caught fish higher in mercury in the previous 60 days, pregnancy status, hair treatment in the previous 30 days, dental filling, and other recent exposures to mercury. To classify commercial consumption versus locally caught fish, respondents were asked one general question about consuming fish higher in mercury in the past 60 days and one question about eating fish higher in mercury that were caught locally. Respondents who answered yes to the first question were classified as consuming commercial fish that is higher in mercury, and if they responded yes to the second question they were classified as consuming locally caught fish that are higher in mercury. Respondents may be counted in both categories. The researchers also looked for a correlation between women’s awareness of mercury contamination and the number of fish meals they consumed per month. The frequency distributions of fish consumption (commercially purchased and locally caught) were also analyzed. The percentages of respondents in the highest and lowest income groups who reported consuming different commercially purchased and locally caught fish types were compared using a two-sample proportion test. The hair mercury concentrations among women who
reported consuming fish higher in mercury (commercially purchased or locally caught) in the previous 60 days and those who did not were compared using the Kruskal–Wallis non-parametric test. The total number of fish meals per month (commercially purchased and locally caught) consumed by women who were aware of the need to limit the consumption of fish high in mercury and those who were not aware was compared using a 2-sample t test. Likewise, the total number of fish meals per month (commercially purchased and locally caught) consumed by women who were aware of local fish advisories and those who were not aware was also compared using a 2-sample t test. The relationship between awareness about the need to limit fish containing higher levels of mercury during pregnancy and awareness about local fish advisories and the consumption of such fish types (locally caught and commercially purchased) were analyzed using Chi square tests. Univariate analyses were performed to identify the variables significantly associated with higher (Cmedian concentration of 0.46 lg/g) hair mercury levels. Univariate associations of categorical variables were analyzed using Chi square/ Fisher’s exact test and those of continuous variables were analyzed using logistic regression. A multivariate logistic regression analysis was performed to understand the association of total fish meal consumption per month with the likelihood of having higher hair mercury levels, controlling for the effect of other variables found to be significantly associated with higher hair mercury levels in univariate analyses.
Results Socio-Demographics A total of 408 women participated in this study; however, due to missing values for some questions, not all demographic categories total to 408. The mean age of the study participants was 36.4 years (SD 8.4 years). On average, the respondents had lived in Martin County for a period of 11.4 years (SD 9.8 years) at the time of the study with a median of 8 years and a range of 1–46 years. Among the women, 18 (4.4 %) reported that they were pregnant at the time of the study, and 147 (32.2 %) reported that they had used a perm or hair relaxer treatment in the previous 30 days. The demographic characteristics of the respondents were comparable to those of the general population in Martin County [19] (Table 1). Hair Mercury Concentration Hair sample results from 408 women were analyzed for mercury. Among these, nine results (seven from White
123
Matern Child Health J Fig. 1 Distribution of hair mercury concentration among respondents in Martin County, Florida
women and two from the ‘Other’ race group) were below the detectable limit. For these nine, the associated mercury concentration was assigned as one half the detectable limits for analysis purposes. The geometric and arithmetic means of hair mercury concentration were 0.371 and 0.676 lg/g of hair, respectively. The median concentration was 0.460 lg/g with a range of 0.008–4.620 lg/g. Figure 1 presents the distribution of hair mercury concentration in the samples collected. Native American women had the highest hair mercury concentration (arithmetic mean 0.925 lg/g, geometric mean 0.428 lg/g, range 0.062–2.030 lg/g, median 0.466 lg/g), followed by White women (arithmetic mean 0.730 lg/g, geometric mean 0.416 lg/g, range 0.008–4.620 lg/g, median 0.518 lg/g). Black women had the lowest hair mercury concentration (arithmetic mean 0.260 lg/g, geometric mean 0.151 lg/g, range 0.020–1.430 lg/g, median 0.146 lg/g). Women in the highest annual household income category had the highest hair mercury concentration (arithmetic mean 0.900 lg/g, geometric mean range 0.574 lg/g, range 0.009–4.620 lg/g, median 0.708 lg/g), and women in the lowest income category had the lowest hair mercury concentration (arithmetic mean 0.390 lg/g, geometric mean 0.220 lg/g, range 0.031–2.320 lg/g, median 0.217 lg/g). However, data on annual household income were not available from all the respondents.
Fish Consumption More than 90 % of the women in the study reported consuming fish. Information was also collected about the species of fish commonly consumed by the respondents. Tuna and grouper were the most widely consumed fish types. Table 2 summarizes the fish consumption pattern on a weekly and monthly basis along with the source of fish (retail store, restaurant and locally caught).
123
The number of different fish meals consumed on a monthly basis along with the source of fish according to age group, race and annual household income categories are presented in Table 3. The mean total fish meal consumption was lower among women under age 25 compared to the older age groups, though not statistically significant. Black women consumed more fish meals than Whites (p B 0.05). A comparison of the percentages of the respondents in the highest (C$75,000) and lowest (\$25,000) income groups who reported ever consuming different commercial fish types indicated that a higher percentage (17 vs 3 %) of women from the higher income group consumed commercially purchased Chilean sea bass (p B 0.05). Though not statistically significant, a higher percentage of women in the higher income group reported ever consuming other commercially available fish such as swordfish (14 vs 5 %) and orange roughy (5 vs 3 %). Likewise, a higher (not statistically significant, p [ 0.05) percentage of women from the higher income group reported ever consuming locally caught grouper (59 vs 39 %) and cobia (10 vs 5 %). Consumption of Fish Known to Contain Higher Levels of Mercury Although more than 90 % of the women reported consuming fish, not all of them reported consuming fish higher in mercury. Among the respondents, 214 (52.5 %) responded consuming either commercially purchased or locally caught fish higher in mercury in the previous 60 days and 94 (23.0 %) reported consuming both commercially purchased and locally caught fish higher in mercury in the same time period. Table 4 presents the characteristics of women who reported consuming fish known to contain higher levels of mercury in the previous 60 days. In addition, seven out of the 18 (38.9 %) pregnant
Matern Child Health J Table 2 Fish consumption pattern and source per week and month among women in Martin County, Florida Item
Mean number of meals
SD
Minimum
Maximum
Median
Meals of canned tuna per week
0.57
0.72
0
4.0
0.3
Meals of canned tuna per month
2.29
2.90
0
16.0
1.0
Meals of canned white tuna per week
0.38
0.65
0
4.0
0.0
Meals of canned white tuna per month
1.51
2.58
0
16.0
0.0
Fish meals at restaurant per week
0.57
0.69
0
5.0
0.5
Fish meals at restaurant per month Fish meals cooked at home per week
2.30 0.67
2.79 0.85
0 0
22.5 7.0
2.0 0.5 2.0
Number of:
Fish meals cooked at home per month
2.65
3.34
0
28.0
Fish meals locally caught per week
0.31
0.56
0
5.0
0.0
Fish meals locally caught per month
1.24
2.24
0
20.0
0.0
Shellfish meals per week
0.69
0.90
0
10.0
0.5
Shellfish meals per month
2.73
3.53
0
40.0
2.0
Table 3 Monthly consumption of fish meals and source of fish (mean ± SD) by participant demographic characteristics in Martin County, Florida Fish meals per month
Canned tuna meals per month
White tuna meals per month
Fish meals at a restaurant per month
Fish meals cooked at home per month
Fish meals locally caught per month
Shellfish meals per month
Total number of fish meals per month
Age group (years)a 18–24
2.37 ± 3.52
1.69 ± 3.34
1.68 ± 2.43
1.33 ± 1.63
1.21 ± 2
2.32 ± 2.83
10.37 ± 10.95
25–34
2.48 ± 3
1.26 ± 2.09
2.21 ± 2.73
2.54 ± 3.06
1.28 ± 2.27
2.81 ± 3.25
12.53 ± 12.11
35–49
2.17 ± 2.7
1.6 ± 2.62
2.47 ± 2.88
2.99 ± 3.68
1.22 ± 2.28
2.77 ± 3.79
13.20 ± 11.38
Raceb White
2.14 ± 2.64
1.52 ± 2.49
2.31 ± 2.71
2.57 ± 3.25
1.17 ± 2.08
2.68 ± 3.44
12.35 ± 10.26
Black
4.55 ± 5.43
2.73 ± 4.61
3.02 ± 4.19
3.25 ± 3.95
2.75 ± 4.1
4.45 ± 5.77
20.75 ± 24.70
Asian
2.29 ± 3.04
1 ± 1.83
1.43 ± 1.81
2±3
0.71 ± 1.11
2 ± 1.73
9.43 ± 7.68
2 ± 1.22
1.6 ± 1.67
1.2 ± 0.84
4 ± 3.54
1.2 ± 0.45
2.2 ± 2.17
12.20 ± 5.63
2.26 ± 2.71
0.52 ± 1.01
1.78 ± 2.71
3.02 ± 4.14
0.93 ± 2.15
2.07 ± 2.32
10.27 ± 10.47
Native American Other Income ($)a \25,000
3.23 ± 3.82
1.64 ± 3.02
1.98 ± 2.99
2.58 ± 3.28
1.37 ± 2.89
2.39 ± 3.59
13.10 ± 15.23
25,000–34,999
2.49 ± 3
1.55 ± 2.68
1.91 ± 1.91
2.79 ± 3.63
1.03 ± 1.63
3.04 ± 3.46
12.81 ± 12.41
35,000–49,999
1.91 ± 2.21
1.13 ± 2
1.82 ± 1.88
2.32 ± 4.33
1.86 ± 2.51
2.07 ± 2.08
11.01 ± 8.25
50,000–74,999
2.07 ± 2.65
1.68 ± 2.71
2.41 ± 2.52
2.27 ± 2.88
1.35 ± 2.66
3.34 ± 5.33
13.07 ± 12.44
2.2 ± 2.64
1.52 ± 2.37
2.89 ± 3.39
3.13 ± 3.33
1.03 ± 1.81
2.82 ± 2.77
13.58 ± 10.13
C75,000 a
No statistically significant difference in the amount of total fish consumed between different subgroups
b
Statistically significant (p B 0.05) differences in the total quantity of fish consumed existed between Whites and Blacks and between Blacks and the ‘Other’ race group
women reported consuming either commercially purchased or locally caught fish higher in mercury in the previous 60 days. Table 5 presents the data on (a) women who were found to have C1 lg of mercury per gram of hair sample, and (b) the characteristics of women with C1 lg of mercury
per gram of hair and who also reported consuming fish higher in mercury in the previous 60 days. Approximately 25 % of the respondents had a mercury concentration C1 lg/g of hair (USEPA reference dose). Twenty-seven percent of White women in the study group had hair mercury concentration C1 lg/g, whereas only 5 % of
123
Matern Child Health J Table 4 Characteristics of women who reported consuming fish higher in mercury in the previous 60 days among women in Martin County, Florida Total sample (n = 408a)
Frequency
Women who consumed fish (commercially purchased or locally caught) higher in mercury in the previous 60 days (n = 214b) Commercially purchased fish (n = 179)
Locally caught fish (n = 129)
Frequency
Frequency
Percent of the stratum
Percent of the stratum
Age (years) 18–24
51
18
35.29
15
29.41
25–34
122
47
38.52
38
31.15
35–49
235
114
48.51
76
32.34
White Black
345 22
158 11
45.8 50
113 9
32.75 40.91
Asian
7
1
14.29
1
14.29
Native American
5
3
60
1
20
28
6
21.43
5
17.86
Race
Other
n = 165a
Income ($)
n = 120a
\25,000
67
21
31.34
15
22.39
25,000–34,999
39
14
35.9
16
41.03
35,000–49,999
53
25
47.17
21
39.62
50,000–74,999
71
34
47.89
29
40.85
136
71
52.21
39
28.68
C75,000 Education Elementary/middle school
12
2
16.67
1
8.33
High school/GED
46
14
30.43
5
10.87
Some college/technical school
119
46
38.66
40
33.61
College graduate/post college
229
117
51.09
83
36.24
a
Due to missing values, not all subgroups total to 408
b
Consumed commercially purchased or locally caught fish higher in mercury
Black women had that level of hair mercury concentration in this study. Women who reported consuming either commercially purchased or locally caught fish higher in mercury in the previous 60 days had higher (p B 0.05) median hair mercury concentration than other women in the study group (0.74 vs 0.26 lg/g). Fish Advisory Awareness The majority (75.3 %) of the respondents were aware of the need to limit consumption of fish higher in mercury during pregnancy, though only 41.5 % were aware of local fish advisories. Though the number of pregnant women in this study was only 18, most of them (83 %) were aware of the need to limit the consumption of fish higher in mercury. However, 56 % of pregnant women were unaware of any local fish advisory. Likewise, a majority of the non-pregnant women were also aware of the need to limit the consumption of fish higher in mercury, but only about 41 % of them were aware of any local fish advisory. The total number of fish
123
meals consumed (commercially purchased and locally caught) per month did not vary (p [ 0.05 for all) by awareness about the need to limit fish higher in mercury during pregnancy or by awareness about local fish advisories. The consumption of fish (commercially purchased and locally caught) was also not related to the awareness of the need to limit the consumption of fish higher in mercury and not related to knowledge of local fish advisories. Factors Associated with Higher Hair Mercury Levels The results of the multivariate logistic regression analysis (presented as odds ratios with corresponding confidence intervals) showing the relationship between different socioeconomic and personal variables with the likelihood of having higher (Cmedian) hair mercury level are presented in Table 6. The likelihood of having higher hair mercury levels increased (p B 0.05) with an increase in the total number of fish meals consumed per month (OR 1.03). Women who did not eat fish (commercially purchased or
Matern Child Health J Table 5 Characteristics of women with a hair mercury C1 lg/g overall and among women who reported consuming fish higher in mercury in the previous 60 days Total sample (n = 408a)
Women with hair mercury level C1 lg/g (n = 100)
Women with hair mercury level C1 lg/g and who also reported consuming fish (commercially purchased or locally caught) higher in mercury in the previous 60 days (n = 77b) Commercially purchased fish (n = 72)
Locally caught fish (n = 45)
Frequency
Frequency
Percent of the stratum
Frequency
Percent of the stratum
Frequency
Percent of the stratum
18–24
51
5
9.8
2
3.92
3
5.88
25–34
122
28
22.95
21
17.21
14
11.48
C35
235
67
28.51
49
20.85
28
11.91
Black
22
1
4.55
1
4.55
1
4.55
Asian
7
0
0
0
0
0
0
Age (years)
Race
Native American
5
2
40
2
40.00
1
20.00
Other
28
5
17.86
3
10.71
3
10.71
White
345
92
26.67
66
19.13
40
a
Income ($)
n = 91
n = 67
a
n = 41
11.59 a
\25,000
67
8
11.94
5
7.46
2
2.99
25,000–34,999 35,000–49,999
39 53
8 11
20.51 20.75
6 10
15.38 18.87
7 7
17.95 13.21
50,000–74,999
71
16
22.54
12
16.90
9
12.68
C75,000
136
48
35.29
34
25.00
16
11.76
12
0
0
0
0
0
Education Elementary/middle school
0
High school/GED
46
6
13.04
3
6.52
0
0
Some college/ technical school College graduate/post college
119
25
21.01
17
14.29
13
10.92
229
69
30.13
52
22.71
32
13.97
a
Due to missing values, not all subgroups total to 408
b
Consumed commercially purchased or locally caught fish higher in mercury
locally caught) known to contain higher levels of mercury during the previous 60 days were approximately 75 % less likely (p B 0.05) to have higher hair mercury levels. Blacks were 81 % less likely (p B 0.05) to have higher hair mercury concentration than Whites. Women with annual household incomes \$25,000 were 62 % less likely to have higher mercury concentration than those with income C$75,000 (p B 0.05).
Discussion Hair Mercury Levels The mean hair mercury levels (geometric mean 0.371 lg/g and arithmetic mean 0.676 lg/g) in women of childbearing
age found in this study were higher than those reported by McDowell et al. [21] of 0.20 and 0.47 lg/g respectively using data from the National Health and Nutrition Examination Survey (NHANES). The arithmetic mean concentration of 0.676 lg/g found in the present study was also higher than that reported based on a small (n = 8), study (0.40 lg/g) in Florida [22]. Another study from Florida [11] reported a median concentration of 0.61 lg/g based on a sample size of 389, which was higher than the 0.460 lg/g found in the present study. The same study reported a median concentration of 0.38 lg/g based on results from 70 participants from Miami, Florida. However, the participants in the Miami study were recruited by various conservation organizations and did not constitute a representative sample of the general population. The higher levels found in the present study compared to the level reported based on the
123
Matern Child Health J Table 6 Odds ratios (OR) and confidence intervals (CI) showing the association between different socioeconomic and personal variables and higher (Cmedian) hair mercury concentration Variable Total fish consumption per month
Beta coefficient
SE
OR (95 % CI)
p Value
0.0297
0.0119
1.030 (1.006, 1.054)
0.013 0.3156
Dental filling Not sure versus yes
-0.5173
0.5155
0.596 (0.217, 1.637)
No versus yes
-0.0777
0.2805
0.925 (0.534, 1.603)
0.7818
Consumption of higher risk fish (no vs yes)
-1.3715
0.2599
0.254 (0.152, 0.422)
Fish Consumption and Hair Mercury Levels in Women of Childbearing Age, Martin County, Florida Anil Nair • Melissa Jordan • Sharon Watkins • Robert Washam • Chris DuClos • Serena Jones • Jason Palcic • Marek Pawlowicz • Carina Blackmore
Ó Springer Science+Business Media New York 2014
Abstract The health effects of mercury in humans are mostly on the developing nervous system. Pregnant women and women who are breastfeeding must be targeted in order to decrease mercury exposure to the populations at highest risk—infants, unborn fetuses, and young children. This purpose of this study is to understand the demographics of fish-consumption patterns among women of childbearing age (including pregnant women) in Martin County, Florida, and to analyze the associations of mercury levels in participants’ hair with socio-demographic variables in order to better design prevention messages and campaigns. Mercury concentrations in hair samples of 408 women ages 18–49 were assessed. Data on demographic factors, pregnancy status, fish consumption, and awareness of fish advisories were collected during personal interviews. Data were analyzed using descriptive statistics and multivariate logistic regression. The geometric and arithmetic means of hair mercury concentration were 0.371 and 0.676 lg/g of hair. One-fourth of the respondents had a concentration C1 lg/g of hair. Consuming a higher number of fish meals per month, consumption of commercially purchased or locally caught fish higher in mercury, White race and income C$75,000 were positively associated with A. Nair M. Jordan (&) S. Watkins C. DuClos C. Blackmore Bureau of Epidemiology, Florida Department of Health, 4052 Bald Cypress Way, Bin A-08, Tallahassee, FL 32399, USA e-mail: [email protected] R. Washam S. Jones Martin County Health Department, Florida Department of Health, 3441 SE Willoughby Boulevard, Stuart, FL 34994, USA J. Palcic M. Pawlowicz Bureau of Public Health Laboratories, Florida Department of Health, 1217 Pearl Street, Jacksonville, FL 32202, USA
the likelihood of having higher hair mercury levels. This study confirms the existence of a higher overall mean hair mercury level and a higher percentage of women with C1 lg/g hair mercury level than those reported at the national level and in other regional studies. This suggests the need for region-specific fish consumption advisories to minimize mercury exposure in humans. Keywords Fish consumption Hair Mercury Women of childbearing age
Introduction Mercury is a metal existing naturally in the earth’s crust. Through a combination of natural and anthropogenic processes, mercury gets deposited in water and becomes methyl mercury through biotransformation by aquatic microorganisms [1, 2]. Fish consumption is the most important pathway of mercury exposure in humans. While all fish have methyl mercury in their tissues, predatory fish that are long-lived tend to have higher body burdens of methyl mercury [1, 3]. The most important toxic effect of methyl mercury is on the development of the nervous system [4]. Methyl mercury can pass through the placental barrier and affect the nervous system of developing fetuses [5]. Developmental neurotoxicity has also been observed following prenatal exposure of children to lower levels of methyl mercury from maternal fish consumption. Studies have reported the adverse developmental effects of high levels of maternal methyl mercury in newborns [6–8]. Using national blood mercury prevalence data from the Centers for Disease Control and Prevention, it was estimated that between 316,588 (7.8 %) and 637,233 (15.8 %) of infants born in the U.S. annually would have cord blood
123
Matern Child Health J
mercury levels [5.8 lg/L [9], a level associated with a decline in IQ. In adults, total blood mercury concentrations reflect methyl mercury exposures over several months, but may be influenced by recent consumption of fish or shellfish meals [10]. Hair sample analysis is a common method used to detect exposure to mercury over several months to a year prior to sampling [11]. Hair concentration is reported to correlate well with dietary methyl mercury intake and hair sample collection is simple and non-invasive [12]. Hair mercury has been used as a biomarker for determining methyl mercury exposure among fish-eating populations [12–15]. Considering the potential adverse effects of methyl mercury, the United States Environmental Protection Agency (USEPA) and the Food and Drug Administration (FDA) have jointly provided fish consumption advisories. Likewise, some states have also provided advisories to limit fish consumption by women of childbearing age. The Florida Department of Health (FDOH) recommends pregnant and nursing women and women who may become pregnant not eat certain fish species, such as: Chilean sea bass, golden snapper, jack, king mackerel, marlin, orange roughy, sea lamprey, shark, Spanish mackerel, sword fish, tile fish, tuna and walleye [16]. According to the World Health Organization [17], dependence on fish as a protein source is higher in coastal than in inland areas. Variation in the mercury content in fish from different regions of Florida has also been reported [18]. The objectives of the present study were to understand the demographic and fish-consumption patterns of women of childbearing age in Martin County, Florida, and to analyze the associations of mercury levels in participants’ hair with socio-demographic variables in order to better design prevention messages and campaigns.
Table 1 Participant demographic characteristics compared with the 2010 U.S. Census Bureau’s reported demographic characteristics for Martin County, Florida [n (%)]
Methods
a
Total sample (n = 408a), women living in Martin County, FL Ages 18–49
Martin County, FL total population = 146,393 U.S. Census Bureau (2010)
Age (years) 18–24
51 (12.5)
10,687 (22.5)
25–34
122 (29.9)
10,979 (23.1)
35–49
235 (57.6)
25,912 (54.5)
345 (84.6) 22 (5.4)
128,510 (88.8) 8,775 (6.0)
Race White Black Asian
7 (1.7)
904 (0.6)
Native American
5 (1.2)
1,765 (1.2)
28 (6.9)
6,439 (4.4)
\25,000
67 (16.4)
12,018 (20.0)
25,000–34,999
39 (9.6)
9,795 (16.3)
35,000–49,999
53 (13.0)
8,953 (14.9)
Other Income ($)b
50,000–74,999
71 (17.4)
9,254 (15.4)
136 (33.3)
20,010 (33.3)
12 (2.9)
14,159 (12.9)
High school/ GED
46 (11.3)
27,989 (25.5)
Some college/ technical school
119 (29.2)
25,136 (22.9)
229 (56.1)
42,478 (38.7)
C75,000 Educationc Elementary/ middle school
College graduate/post college
Due to missing values, not all subgroups total to 408
b
Human Subjects Review
Census data for income are reported by number of households at each level, not individuals
c
The study protocol was submitted to the FDOH Ethics and Human Research Protection Program in April 2009 to determine if this mercury biomonitoring activity involved research. It was determined that this study was designed to establish baseline mercury levels for purpose of public health surveillance and is not primarily an activity designed to contribute to generalizable knowledge, so review by the FDOH Institutional Review Board was not required.
Census data on educational attainment are only available for those ages 25 and older
had resided in Martin County, Florida for at least 1 year. A recruitment announcement for women was posted in newspapers, physician offices, health clubs, health department clinics and websites, and advertised on local public television stations. Although not a random sampling, the final demographic characteristics of respondents were not markedly different from the overall county population [19] (Table 1).
Sample Selection Data and Hair Sample Collection We performed a cross-sectional study in a medium sized coastal county in Florida. The data for this study were collected in 2010 from 408 volunteer women ages 18–49, who
123
Upon verification of the eligibility criteria, the selected volunteers were asked to sign an informed consent form
Matern Child Health J
outlining the details of the project. Data on a variety of factors, including: fish consumption habits, knowledge of fish consumption advisories, pregnancy status, demographic and socio-economic information and mercury exposure history were collected from the selected respondents using a pre-tested interview form. The interview questions had previously been used in two related studies in Florida. At the end of the interview, the respondents were asked to provide approximately 50 mg (*100 strands) of their hair sample cut close to the scalp from the occipital region of the head using stainless steel surgical scissors. Hair samples were analyzed for concentration of total mercury by the FDOH Bureau of Public Health Laboratories in Jacksonville by combustion-gold amalgamation-atomic absorption spectroscopy (C-GA-AAS) with the MA-2000 mercury analyzer (Nippon Instruments Corporation, Tokyo, Japan). The calibration linear dynamic range was modified (2- to 60-ng of mercury) to achieve lower detection limits than previously reported [20]. Statistical Analysis All analyses were performed using SAS software (version 9.2, SAS Institute, Carey, NC) and a p value of B0.05 was considered significant in all analyses. Descriptive statistics were used to present the frequency distribution of the number of participants according to age group, race, educational attainment, annual household income, possession of a fishing license, fish consumption patterns, awareness about limiting fish consumption, awareness about local fish advisories, consumption of commercially purchased or locally caught fish higher in mercury in the previous 60 days, pregnancy status, hair treatment in the previous 30 days, dental filling, and other recent exposures to mercury. To classify commercial consumption versus locally caught fish, respondents were asked one general question about consuming fish higher in mercury in the past 60 days and one question about eating fish higher in mercury that were caught locally. Respondents who answered yes to the first question were classified as consuming commercial fish that is higher in mercury, and if they responded yes to the second question they were classified as consuming locally caught fish that are higher in mercury. Respondents may be counted in both categories. The researchers also looked for a correlation between women’s awareness of mercury contamination and the number of fish meals they consumed per month. The frequency distributions of fish consumption (commercially purchased and locally caught) were also analyzed. The percentages of respondents in the highest and lowest income groups who reported consuming different commercially purchased and locally caught fish types were compared using a two-sample proportion test. The hair mercury concentrations among women who
reported consuming fish higher in mercury (commercially purchased or locally caught) in the previous 60 days and those who did not were compared using the Kruskal–Wallis non-parametric test. The total number of fish meals per month (commercially purchased and locally caught) consumed by women who were aware of the need to limit the consumption of fish high in mercury and those who were not aware was compared using a 2-sample t test. Likewise, the total number of fish meals per month (commercially purchased and locally caught) consumed by women who were aware of local fish advisories and those who were not aware was also compared using a 2-sample t test. The relationship between awareness about the need to limit fish containing higher levels of mercury during pregnancy and awareness about local fish advisories and the consumption of such fish types (locally caught and commercially purchased) were analyzed using Chi square tests. Univariate analyses were performed to identify the variables significantly associated with higher (Cmedian concentration of 0.46 lg/g) hair mercury levels. Univariate associations of categorical variables were analyzed using Chi square/ Fisher’s exact test and those of continuous variables were analyzed using logistic regression. A multivariate logistic regression analysis was performed to understand the association of total fish meal consumption per month with the likelihood of having higher hair mercury levels, controlling for the effect of other variables found to be significantly associated with higher hair mercury levels in univariate analyses.
Results Socio-Demographics A total of 408 women participated in this study; however, due to missing values for some questions, not all demographic categories total to 408. The mean age of the study participants was 36.4 years (SD 8.4 years). On average, the respondents had lived in Martin County for a period of 11.4 years (SD 9.8 years) at the time of the study with a median of 8 years and a range of 1–46 years. Among the women, 18 (4.4 %) reported that they were pregnant at the time of the study, and 147 (32.2 %) reported that they had used a perm or hair relaxer treatment in the previous 30 days. The demographic characteristics of the respondents were comparable to those of the general population in Martin County [19] (Table 1). Hair Mercury Concentration Hair sample results from 408 women were analyzed for mercury. Among these, nine results (seven from White
123
Matern Child Health J Fig. 1 Distribution of hair mercury concentration among respondents in Martin County, Florida
women and two from the ‘Other’ race group) were below the detectable limit. For these nine, the associated mercury concentration was assigned as one half the detectable limits for analysis purposes. The geometric and arithmetic means of hair mercury concentration were 0.371 and 0.676 lg/g of hair, respectively. The median concentration was 0.460 lg/g with a range of 0.008–4.620 lg/g. Figure 1 presents the distribution of hair mercury concentration in the samples collected. Native American women had the highest hair mercury concentration (arithmetic mean 0.925 lg/g, geometric mean 0.428 lg/g, range 0.062–2.030 lg/g, median 0.466 lg/g), followed by White women (arithmetic mean 0.730 lg/g, geometric mean 0.416 lg/g, range 0.008–4.620 lg/g, median 0.518 lg/g). Black women had the lowest hair mercury concentration (arithmetic mean 0.260 lg/g, geometric mean 0.151 lg/g, range 0.020–1.430 lg/g, median 0.146 lg/g). Women in the highest annual household income category had the highest hair mercury concentration (arithmetic mean 0.900 lg/g, geometric mean range 0.574 lg/g, range 0.009–4.620 lg/g, median 0.708 lg/g), and women in the lowest income category had the lowest hair mercury concentration (arithmetic mean 0.390 lg/g, geometric mean 0.220 lg/g, range 0.031–2.320 lg/g, median 0.217 lg/g). However, data on annual household income were not available from all the respondents.
Fish Consumption More than 90 % of the women in the study reported consuming fish. Information was also collected about the species of fish commonly consumed by the respondents. Tuna and grouper were the most widely consumed fish types. Table 2 summarizes the fish consumption pattern on a weekly and monthly basis along with the source of fish (retail store, restaurant and locally caught).
123
The number of different fish meals consumed on a monthly basis along with the source of fish according to age group, race and annual household income categories are presented in Table 3. The mean total fish meal consumption was lower among women under age 25 compared to the older age groups, though not statistically significant. Black women consumed more fish meals than Whites (p B 0.05). A comparison of the percentages of the respondents in the highest (C$75,000) and lowest (\$25,000) income groups who reported ever consuming different commercial fish types indicated that a higher percentage (17 vs 3 %) of women from the higher income group consumed commercially purchased Chilean sea bass (p B 0.05). Though not statistically significant, a higher percentage of women in the higher income group reported ever consuming other commercially available fish such as swordfish (14 vs 5 %) and orange roughy (5 vs 3 %). Likewise, a higher (not statistically significant, p [ 0.05) percentage of women from the higher income group reported ever consuming locally caught grouper (59 vs 39 %) and cobia (10 vs 5 %). Consumption of Fish Known to Contain Higher Levels of Mercury Although more than 90 % of the women reported consuming fish, not all of them reported consuming fish higher in mercury. Among the respondents, 214 (52.5 %) responded consuming either commercially purchased or locally caught fish higher in mercury in the previous 60 days and 94 (23.0 %) reported consuming both commercially purchased and locally caught fish higher in mercury in the same time period. Table 4 presents the characteristics of women who reported consuming fish known to contain higher levels of mercury in the previous 60 days. In addition, seven out of the 18 (38.9 %) pregnant
Matern Child Health J Table 2 Fish consumption pattern and source per week and month among women in Martin County, Florida Item
Mean number of meals
SD
Minimum
Maximum
Median
Meals of canned tuna per week
0.57
0.72
0
4.0
0.3
Meals of canned tuna per month
2.29
2.90
0
16.0
1.0
Meals of canned white tuna per week
0.38
0.65
0
4.0
0.0
Meals of canned white tuna per month
1.51
2.58
0
16.0
0.0
Fish meals at restaurant per week
0.57
0.69
0
5.0
0.5
Fish meals at restaurant per month Fish meals cooked at home per week
2.30 0.67
2.79 0.85
0 0
22.5 7.0
2.0 0.5 2.0
Number of:
Fish meals cooked at home per month
2.65
3.34
0
28.0
Fish meals locally caught per week
0.31
0.56
0
5.0
0.0
Fish meals locally caught per month
1.24
2.24
0
20.0
0.0
Shellfish meals per week
0.69
0.90
0
10.0
0.5
Shellfish meals per month
2.73
3.53
0
40.0
2.0
Table 3 Monthly consumption of fish meals and source of fish (mean ± SD) by participant demographic characteristics in Martin County, Florida Fish meals per month
Canned tuna meals per month
White tuna meals per month
Fish meals at a restaurant per month
Fish meals cooked at home per month
Fish meals locally caught per month
Shellfish meals per month
Total number of fish meals per month
Age group (years)a 18–24
2.37 ± 3.52
1.69 ± 3.34
1.68 ± 2.43
1.33 ± 1.63
1.21 ± 2
2.32 ± 2.83
10.37 ± 10.95
25–34
2.48 ± 3
1.26 ± 2.09
2.21 ± 2.73
2.54 ± 3.06
1.28 ± 2.27
2.81 ± 3.25
12.53 ± 12.11
35–49
2.17 ± 2.7
1.6 ± 2.62
2.47 ± 2.88
2.99 ± 3.68
1.22 ± 2.28
2.77 ± 3.79
13.20 ± 11.38
Raceb White
2.14 ± 2.64
1.52 ± 2.49
2.31 ± 2.71
2.57 ± 3.25
1.17 ± 2.08
2.68 ± 3.44
12.35 ± 10.26
Black
4.55 ± 5.43
2.73 ± 4.61
3.02 ± 4.19
3.25 ± 3.95
2.75 ± 4.1
4.45 ± 5.77
20.75 ± 24.70
Asian
2.29 ± 3.04
1 ± 1.83
1.43 ± 1.81
2±3
0.71 ± 1.11
2 ± 1.73
9.43 ± 7.68
2 ± 1.22
1.6 ± 1.67
1.2 ± 0.84
4 ± 3.54
1.2 ± 0.45
2.2 ± 2.17
12.20 ± 5.63
2.26 ± 2.71
0.52 ± 1.01
1.78 ± 2.71
3.02 ± 4.14
0.93 ± 2.15
2.07 ± 2.32
10.27 ± 10.47
Native American Other Income ($)a \25,000
3.23 ± 3.82
1.64 ± 3.02
1.98 ± 2.99
2.58 ± 3.28
1.37 ± 2.89
2.39 ± 3.59
13.10 ± 15.23
25,000–34,999
2.49 ± 3
1.55 ± 2.68
1.91 ± 1.91
2.79 ± 3.63
1.03 ± 1.63
3.04 ± 3.46
12.81 ± 12.41
35,000–49,999
1.91 ± 2.21
1.13 ± 2
1.82 ± 1.88
2.32 ± 4.33
1.86 ± 2.51
2.07 ± 2.08
11.01 ± 8.25
50,000–74,999
2.07 ± 2.65
1.68 ± 2.71
2.41 ± 2.52
2.27 ± 2.88
1.35 ± 2.66
3.34 ± 5.33
13.07 ± 12.44
2.2 ± 2.64
1.52 ± 2.37
2.89 ± 3.39
3.13 ± 3.33
1.03 ± 1.81
2.82 ± 2.77
13.58 ± 10.13
C75,000 a
No statistically significant difference in the amount of total fish consumed between different subgroups
b
Statistically significant (p B 0.05) differences in the total quantity of fish consumed existed between Whites and Blacks and between Blacks and the ‘Other’ race group
women reported consuming either commercially purchased or locally caught fish higher in mercury in the previous 60 days. Table 5 presents the data on (a) women who were found to have C1 lg of mercury per gram of hair sample, and (b) the characteristics of women with C1 lg of mercury
per gram of hair and who also reported consuming fish higher in mercury in the previous 60 days. Approximately 25 % of the respondents had a mercury concentration C1 lg/g of hair (USEPA reference dose). Twenty-seven percent of White women in the study group had hair mercury concentration C1 lg/g, whereas only 5 % of
123
Matern Child Health J Table 4 Characteristics of women who reported consuming fish higher in mercury in the previous 60 days among women in Martin County, Florida Total sample (n = 408a)
Frequency
Women who consumed fish (commercially purchased or locally caught) higher in mercury in the previous 60 days (n = 214b) Commercially purchased fish (n = 179)
Locally caught fish (n = 129)
Frequency
Frequency
Percent of the stratum
Percent of the stratum
Age (years) 18–24
51
18
35.29
15
29.41
25–34
122
47
38.52
38
31.15
35–49
235
114
48.51
76
32.34
White Black
345 22
158 11
45.8 50
113 9
32.75 40.91
Asian
7
1
14.29
1
14.29
Native American
5
3
60
1
20
28
6
21.43
5
17.86
Race
Other
n = 165a
Income ($)
n = 120a
\25,000
67
21
31.34
15
22.39
25,000–34,999
39
14
35.9
16
41.03
35,000–49,999
53
25
47.17
21
39.62
50,000–74,999
71
34
47.89
29
40.85
136
71
52.21
39
28.68
C75,000 Education Elementary/middle school
12
2
16.67
1
8.33
High school/GED
46
14
30.43
5
10.87
Some college/technical school
119
46
38.66
40
33.61
College graduate/post college
229
117
51.09
83
36.24
a
Due to missing values, not all subgroups total to 408
b
Consumed commercially purchased or locally caught fish higher in mercury
Black women had that level of hair mercury concentration in this study. Women who reported consuming either commercially purchased or locally caught fish higher in mercury in the previous 60 days had higher (p B 0.05) median hair mercury concentration than other women in the study group (0.74 vs 0.26 lg/g). Fish Advisory Awareness The majority (75.3 %) of the respondents were aware of the need to limit consumption of fish higher in mercury during pregnancy, though only 41.5 % were aware of local fish advisories. Though the number of pregnant women in this study was only 18, most of them (83 %) were aware of the need to limit the consumption of fish higher in mercury. However, 56 % of pregnant women were unaware of any local fish advisory. Likewise, a majority of the non-pregnant women were also aware of the need to limit the consumption of fish higher in mercury, but only about 41 % of them were aware of any local fish advisory. The total number of fish
123
meals consumed (commercially purchased and locally caught) per month did not vary (p [ 0.05 for all) by awareness about the need to limit fish higher in mercury during pregnancy or by awareness about local fish advisories. The consumption of fish (commercially purchased and locally caught) was also not related to the awareness of the need to limit the consumption of fish higher in mercury and not related to knowledge of local fish advisories. Factors Associated with Higher Hair Mercury Levels The results of the multivariate logistic regression analysis (presented as odds ratios with corresponding confidence intervals) showing the relationship between different socioeconomic and personal variables with the likelihood of having higher (Cmedian) hair mercury level are presented in Table 6. The likelihood of having higher hair mercury levels increased (p B 0.05) with an increase in the total number of fish meals consumed per month (OR 1.03). Women who did not eat fish (commercially purchased or
Matern Child Health J Table 5 Characteristics of women with a hair mercury C1 lg/g overall and among women who reported consuming fish higher in mercury in the previous 60 days Total sample (n = 408a)
Women with hair mercury level C1 lg/g (n = 100)
Women with hair mercury level C1 lg/g and who also reported consuming fish (commercially purchased or locally caught) higher in mercury in the previous 60 days (n = 77b) Commercially purchased fish (n = 72)
Locally caught fish (n = 45)
Frequency
Frequency
Percent of the stratum
Frequency
Percent of the stratum
Frequency
Percent of the stratum
18–24
51
5
9.8
2
3.92
3
5.88
25–34
122
28
22.95
21
17.21
14
11.48
C35
235
67
28.51
49
20.85
28
11.91
Black
22
1
4.55
1
4.55
1
4.55
Asian
7
0
0
0
0
0
0
Age (years)
Race
Native American
5
2
40
2
40.00
1
20.00
Other
28
5
17.86
3
10.71
3
10.71
White
345
92
26.67
66
19.13
40
a
Income ($)
n = 91
n = 67
a
n = 41
11.59 a
\25,000
67
8
11.94
5
7.46
2
2.99
25,000–34,999 35,000–49,999
39 53
8 11
20.51 20.75
6 10
15.38 18.87
7 7
17.95 13.21
50,000–74,999
71
16
22.54
12
16.90
9
12.68
C75,000
136
48
35.29
34
25.00
16
11.76
12
0
0
0
0
0
Education Elementary/middle school
0
High school/GED
46
6
13.04
3
6.52
0
0
Some college/ technical school College graduate/post college
119
25
21.01
17
14.29
13
10.92
229
69
30.13
52
22.71
32
13.97
a
Due to missing values, not all subgroups total to 408
b
Consumed commercially purchased or locally caught fish higher in mercury
locally caught) known to contain higher levels of mercury during the previous 60 days were approximately 75 % less likely (p B 0.05) to have higher hair mercury levels. Blacks were 81 % less likely (p B 0.05) to have higher hair mercury concentration than Whites. Women with annual household incomes \$25,000 were 62 % less likely to have higher mercury concentration than those with income C$75,000 (p B 0.05).
Discussion Hair Mercury Levels The mean hair mercury levels (geometric mean 0.371 lg/g and arithmetic mean 0.676 lg/g) in women of childbearing
age found in this study were higher than those reported by McDowell et al. [21] of 0.20 and 0.47 lg/g respectively using data from the National Health and Nutrition Examination Survey (NHANES). The arithmetic mean concentration of 0.676 lg/g found in the present study was also higher than that reported based on a small (n = 8), study (0.40 lg/g) in Florida [22]. Another study from Florida [11] reported a median concentration of 0.61 lg/g based on a sample size of 389, which was higher than the 0.460 lg/g found in the present study. The same study reported a median concentration of 0.38 lg/g based on results from 70 participants from Miami, Florida. However, the participants in the Miami study were recruited by various conservation organizations and did not constitute a representative sample of the general population. The higher levels found in the present study compared to the level reported based on the
123
Matern Child Health J Table 6 Odds ratios (OR) and confidence intervals (CI) showing the association between different socioeconomic and personal variables and higher (Cmedian) hair mercury concentration Variable Total fish consumption per month
Beta coefficient
SE
OR (95 % CI)
p Value
0.0297
0.0119
1.030 (1.006, 1.054)
0.013 0.3156
Dental filling Not sure versus yes
-0.5173
0.5155
0.596 (0.217, 1.637)
No versus yes
-0.0777
0.2805
0.925 (0.534, 1.603)
0.7818
Consumption of higher risk fish (no vs yes)
-1.3715
0.2599
0.254 (0.152, 0.422)
