Blood lead concentrations in the United States have declined dramatically since the 1970s because of the phaseout of leaded gasoline, the ban of lead in paint and consumer products, and the discontinuation of lead use in plumbing and domestically manufactured soldered cans (Annest et al. 1983; Brody et al. 1994; Muntner et al. 2003; Pirkle et al. 1994). However, even at current lower levels, evidence suggests that pre- or postnatal exposure can potentially impair a child’s intellectual function (Baghurst et al. 1992; Canfield et al. 2003; Gomaa et al. 2002; Lanphear et al. 2005; Schnaas et al. 2006). Low-to-moderate levels of lead exposure in pregnancy may also increase the risk of spontaneous abortion (Borja-Aburto et al. 1999) and preterm birth (Andrews et al. 1994). In the general adult population, lead exposure has been associated with elevated blood pressure and hypertension (Martin et al. 2006; Nash et al. 2003), kidney disease (Kim et al. 1996), peripheral arterial disease (Muntner et al. 2003; Navas-Acien et al. 2004), and cardiovascular and all-cause mortality (Menke et al. 2006; Schober et al. 2006).

Cadmium occurs naturally in some soils in addition to being deposited through emissions from mining operations and fossil fuel combustion, application of phosphate fertilizer or sewage sludge, and disposal of cadmium-containing products [Agency for Toxic Substances and Disease Registry (ATSDR) 1999]. Tobacco and food crops can take up cadmium from the soil, and shellfish can accumulate cadmium from the aquatic environment, making cigarette smoke and diet the principal sources of nonoccupational exposure in the United States. Epidemiologic evidence has linked relatively low-level cadmium exposure to renal dysfunction (Buchet et al. 1990; Jarup and Alfven 2004) and decreased bone mineral density (Åkesson et al. 2006; Staessen et al. 1999).

Exposure to mercury in the United States occurs predominantly from consumption of predatory fish that have bioaccumulated methylmercury from the aquatic environment (Bjornberg et al. 2003; Sanzo et al. 2001; Svensson et al. 1992). Methylmercury can cross the blood–brain barrier and interfere with functioning of the central nervous system. Children’s developing nervous systems appear to be most vulnerable [National Research Council (NRC) 2000]. Deficits in language, attention, and memory among children exposed in utero have been reported in studies from the Faroe Islands, New Zealand, and the United States (Crump et al. 1998; Grandjean et al. 1999; Oken et al. 2005). Because methylmercury easily crosses the placenta (Ask et al. 2002; Vahter et al. 2000) and concentrates in fetal blood (Stern and Smith 2003), exposure in women of reproductive age is of particular concern. There is also evidence that methylmercury exposure in adulthood might interfere with vision, motor function, and memory (Lebel et al. 1998; Yokoo et al. 2003) as well as increase the risk of cardiovascular disease (Virtanen et al. 2007).

Ongoing surveillance of exposure to toxic substances is essential for identifying and targeting high-risk groups, evaluating interventions, tracking exposure over time, and monitoring exposures during emergency situations. New York State (NYS) law requires that all children be tested for lead at 1 and 2 years of age. NYS law also requires clinical laboratories to report all blood lead levels and elevated levels of mercury and cadmium in blood or urine to the State Heavy Metals Registry. However, testing among adults is voluntary and, therefore, likely to overrepresent higher-risk groups, for example, those in certain occupations or who request tests because of known or suspected exposures.

In 2004 New York City (NYC) conducted the first-ever local health and nutrition examination survey [NYC Health and Nutrition Examination Survey (HANES)] in a representative sample of NYC adults. The survey measured blood concentrations of lead, mercury, and cadmium using a design that mirrored the National Health and Nutrition Examination Survey (NHANES). In the present article we describe blood metal concentrations by demographic and behavioral characteristics. Results will be used to prioritize public health actions in NYC, where demographics and environment differ in many respects from the United States as a whole.

We describe blood lead results in Table 1. Geometric mean blood lead concentrations increased with age and decreased with income, education, and length of residence in the United States for the foreign-born (p-values for trend tests < 0.04). Blood lead concentrations were highest in heavy smokers (2.49 μg/dL), the foreign-born Chinese (2.66 μg/dL), and those working in construction and maintenance (2.86 μg/dL). Upon removal of the latter group, the geometric mean blood lead level in smokers decreased slightly to 2.00 μg/dL (95th percentile = 5.51 μg/dL), suggesting some confounding of the smoking association by occupation. Prevalence of current smoking among construction and maintenance workers was 45% compared with a citywide estimate of 23%.

The patterns of lead concentrations across population subgroups were similar after we adjusted for predictors simultaneously in a log-linear regression—with several exceptions. The crude association between decreasing income and increasing geometric mean blood lead was no longer apparent (p-value for trend test = 0.54), and former smokers had only 8% higher blood lead concentrations than never smokers (compared with a crude elevation of 26%). Age remained the strongest predictor of blood lead. Upon adding blood cadmium to the adjusted model, a 1-μg/L increase predicted a 22% elevation (95% CI, 17–28%) in mean blood lead concentration.

The geometric mean blood cadmium concentration in NYC adults was 0.77 μg/L (95% CI, 0.75–0.80) as shown in Table 2. All sample levels exceeded the limit of detection and ranged from 0.25 to 9.67 μg/L. There were four blood cadmium levels > 5 μg/L, two of which were measured in foreign-born Chinese males who also had blood concentrations of mercury > 15 μg/L (n = 1) or lead > 10 μg/dL (n = 1). No samples attained the NYS reportable level for cadmium of 10 μg/L, although the highest measured level of 9.67 μg/L came close. The 97.5th percentile for blood cadmium concentration overall was 2.49 μg/L.

Blood cadmium levels were most strongly associated with smoking status. Heavy smokers had the highest geometric mean cadmium concentration (1.58 μg/L) of all subgroups examined. However, the geometric mean among foreign-born Chinese New Yorkers (1.34 μg/L) exceeded that of current smokers (1.22 μg/L), even though the estimated prevalence of smoking in this population subgroup (21%) was not higher than that of the general adult population (24%).

Results from a multiple linear regression were consistent with the patterns of crude geometric means observed across population subgroups. Current smoking and Asian race/ethnicity remained the strongest predictors of elevated blood cadmium. Blood lead was a relatively strong predictor of blood cadmium. After adjusting for other predictors, a 5-μg/dL increase in blood lead concentration predicted a 17% elevation in blood cadmium concentration (95% CI, 5%–31%).

The geometric mean blood mercury concentration among NYC adults was 2.73 μg/L (95% CI, 2.58–2.89) as shown in Table 3. All sample values exceeded the limit of detection and ranged between 0.21 and 35.78 μg/L. About one quarter (24.8%; 95% CI, 22.2–27.7%), or 1.4 million NYC adults, had blood mercury concentrations equaling or exceeding the NYS reportable level of 5 μg/L. There were 54 participants (population prevalence = 2.8%) who exceeded the NYS investigation level of 15 μg/L. Women 20–49 years of age had a geometric mean blood mercury level of 2.64 μg/L and a 23.8% prevalence of blood mercury ≥5 μg/L, similar to the total population. The 97.5th percentile for blood mercury concentration overall was 15.37 μg/L

Frequent consumption of fish or shellfish was associated with increasing mercury levels (p-values for trend test < 0.01 for geometric mean and prevalence ≥5 μg/L) (Table 3). The geometric mean blood level in those who reported consuming fish or shellfish 20 times or more in the last 30 days (5.65 μg/L) was more than 4 times the level of those who did not consume fish or shell-fish (1.31 μg/L). Over half (56.2%) of those who reported consuming fish 20 or more times in the last 30 days had mercury levels ≥5 μg/L, almost eight times the prevalence in those who did not consume fish or shellfish (7.3%).

People born outside the United States had higher mercury levels than those born in the United States; however we did not see a trend toward increasing mercury concentration with shorter time in the United States as we did with lead levels. In contrast, those who had lived in the United States for > 10 years had a higher crude geometric mean blood mercury level than newer arrivals (p < 0.01).

The geometric mean blood mercury level in Asians was higher than other racial/ethnic groups (4.11 μg/L). The 95th percentile was among the highest (19.19 μg/L). The geometric mean in foreign-born Chinese New Yorkers was even higher (7.26 μg/L), surpassing that of all other subgroups we examined. Almost half of adult Asian New Yorkers (46.2%) had blood mercury ≥5 g/L. Among the 93 foreign-born Chinese New Yorkers in the survey, 68 had blood mercury concentrations ≥5 μg/L (population prevalence = 71.7%), and 19 of these were ≥15 μg/L (population prevalence = 20.0%).

NHANES 1999–2002 (CDC 2005a) provided national estimates of blood mercury concentration for women 16–49 years of age. The geometric mean blood mercury concentration in our slightly older sample of NYC women 20–49 years of age (2.64 μg/L) is more than 3 times the NHANES 2001–2002 estimate (0.83 μg/L) [Centers for Disease Control (CDC) 2005a; Figure 1]. This elevation is consistent with a previous report of higher blood mercury levels in the Eastern coastal region of the United States relative to the United States as a whole (Mahaffey 2005).

Blood mercury levels were higher in NYC than nationally across similar levels of reported fish or shellfish consumption (Mahaffey et al. 2004). A possible explanation for this observation is that New Yorkers consume more heavily contaminated fish. A similar scenario may be occurring in the higher income groups, where mercury levels remain elevated even after adjustment for frequency of fish or shellfish consumption. Elevations in economically advantaged individuals may be due to consumption of more expensive fish, such as swordfish, which tend to be higher in mercury (Hightower and Moore 2003). However, even comparing people who reported no fish or shellfish consumption in the past 30 days, the geometric mean blood mercury concentration among New Yorkers was 3 times the national level (Mahaffey et al. 2004).

Blood metal concentrations among Asians have not routinely been reported from the NHANES because of sample size limitations. However, an analysis of 1999–2002 data identified the aggregate of Asians, Pacific Islanders, Native Americans and multiracial groups as having the highest mercury levels of all race/ethnicities (Hightower et al. 2006), similar to our findings. In NYC, fish consumption is the most likely explanation for the racial and ethnic differences in mercury exposure; consumption of at least 20 meals of fish or seafood in the last 30 days was highest in Asians (19%) compared with Whites or Blacks (5.5% each) and Hispanics (1.3%).

We are not aware of NHANES reports that describe elevated blood cadmium or lead in Asians, either alone or as an aggregate group, so we do not know whether the higher levels we measured among Asian New Yorkers mirror national data. Current smoking did not explain the higher cadmium or lead levels in Asians; in fact, prevalence of current smoking was slightly lower among Asian New Yorkers compared with the citywide estimate. Shellfish consumption is a possible source of the higher cadmium levels observed in Asians. Exposure could have occurred outside the United States as well, as cadmium and lead can remain in the body for decades, and body stores may serve as a source of subsequently measured metals in blood (Gulson et al. 1995; Nordberg and Kjellstrom 1979; Smith et al. 1996). In NYC, a large percentage (92%) of Asian adults are foreign-born (U.S. Census Bureau 2000).

The geometric mean blood lead concentration in NYC adults (1.79 μg/dL) is similar to the 2001–2002 national estimate (1.56 μg/dL; CDC 2005a; Figure 1). Despite declining trends (Muntner et al. 2005), current exposure levels have been associated with adverse health effects in children and adults (Canfield et al. 2003; Menke et al. 2006). In adults, nonoccupational lead exposure can occur during renovation of homes or other structures that used lead-based paints in the past. Residential remodeling was the likely source of exposure for the largest number of nonoccupational cases of blood lead ≥25 μg/dL reported to the NYS Heavy Metals Registry 2000–2005 (New York State Department of Health 2006). Other exposure sources included target shooting, ingestion (pica), lead-glazed pottery, soil, dust, and some imported food, spices and traditional medicines (ATSDR 2005; CDC 2005b; Saper et al. 2004). Cigarette smoke contains only small amounts of lead (ATSDR 2005), but our results are consistent with previous reports of positive associations between passive and active smoking and blood lead (Mannino et al. 2005; Shaper et al. 1982). It is possible that the association we observed was confounded by occupational lead exposure, as lead levels among current smokers decrease upon exclusion of persons who reported working in construction or maintenance.

The geometric mean blood cadmium concentration in NYC adults (0.77 μg/L) is slightly higher than the 1999–2000 national estimate for adults (0.47 μg/L; CDC 2005a; Figure 1). Though the difference appears to be statistically significant (judging from the nonoverlapping confidence intervals), the clinical or biological significance of a 0.3 μg/L elevation is not known. Decreased bone mineral density in older women has been associated with blood cadmium levels ≥1.1 μg/L (Alfven et al. 2000), which are typical of current smokers and the foreign-born Chinese in our survey. Cadmium is a constituent of cigarette smoke (ATSDR 1999), and the strong association between current smoking and blood cadmium provides further motivation to prevent smoking initiation and to promote smoking cessation.

Our findings have some limitations. Although the sample selection was designed to be representative of the NYC adult population, we cannot rule out the presence of bias, as the overall response rate was 50%. However, to correct for bias, sample weights incorporated information on age, sex, race/ethnicity, income, education, language spoken at home, and household size, obtained either directly from interview or from neighborhood census data. We also note that the NHANES interview and examination response rate for a similarly aged population in the NYC area in 2004 was only slightly higher, 58% (personal communication with the NHANES program), compared with the 55% response in the NYC HANES (response rates for blood collection component of the examination are slightly lower in both surveys).

Self-reported exposure data are limited by respondents’ memories and ability to answer questions. We do not know how accurately respondents were able to provide the number of times they ate fish or shellfish in the last 30 days. Furthermore, our questionnaire did not distinguish consumption of fish species according to mercury content. Consequently, confounding by contaminated fish and seafood consumption is likely to remain in our comparisons of mercury levels across population subgroups after adjustment for fish or shellfish consumption.

Laboratory methods for determining chemical exposures have become increasingly sensitive, so the detection of lead, mercury or cadmium in the blood of an adult does not necessarily imply a health risk. Findings are difficult to interpret in terms of public health impact, as reference doses are not necessarily meaningful threshold values for toxicity. The data we present attempt to describe exposures in the NYC adult population for the purpose of targeting intervention to high-risk groups and establishing baseline exposure levels.