Modification by ALAD of the Association between Blood Lead and Blood Pressure in the U.S. Population: Results from the Third National Health and Nutrition Examination Survey

Background Environmental lead exposure has been found to be associated with an increased risk of hypertension. Individuals vary greatly in susceptibility to lead toxicity, and genetic susceptibility has often been cited as the probable cause for such variation. Objective The main objective is to determine the role of the aminolevulinic acid dehydratase (ALAD) gene, which encodes the main carrier protein of lead in blood, in the association between lead exposure and blood pressure (BP) and hypertension in the U.S. population. Methods We analyzed data from individuals ≥ 17 years of age who participated in the Third National Health and Nutrition Examination Survey for whom DNA was available (n = 6,016). Multivariable logistic and linear regressions stratified by race/ethnicity were used to examine whether hypertension and BP were associated with ALAD and blood lead levels (BLL). Results BLL was associated with systolic BP in non-Hispanic whites and with hypertension and systolic and diastolic BP in non-Hispanic blacks. BLL was not associated with BP outcomes in Mexican Americans. Non-Hispanic white ALAD2 carriers in the highest BLL quartile (3.8–52.9 μg/dL) had a significantly higher adjusted prevalence odds ratio for hypertension compared with ALAD1 homozygous individuals. We also found a significant interaction between lead concentration and the ALAD2 allele in non-Hispanic whites and non-Hispanic blacks in relation to systolic BP. Conclusions BLL may be an important risk factor for hypertension and increased systolic and diastolic BP. These associations may be modified by ALAD genotype.

Occupational studies have repeatedly demon strated that blood lead levels (BLL) > 40 µg/dL are associated with increased risk of elevated blood pressure (BP) (reviewed by NavasAcien et al. 2007). A recent metaanalysis reported that a 2fold increase in BLL was associated with a 1.0mmHg and a 0.6mmHg increase in systolic and diastolic BP, respectively (Nawrot et al. 2002). Previous studies have demonstrated associations between lowlevel lead exposure and increased BP Nash et al. 2003;NavasAcien et al. 2004;Nordberg et al. 2000;Pirkle et al. 1985;Vupputuri et al. 2003), but the results have been uneven, suggesting that other bio logical factors (e.g., genetic and race/ethnicity) may also be operating.
The δaminolevulinic acid dehydratase (ALAD) enzyme catalyzes the second step in heme biosynthesis and is known to be the major carrier protein for lead in blood (Bergdahl et al. 1997). ALAD, which in humans is encoded by a single gene local ized to the chromosome 9q34 region, is a polymorphic enzyme with two codomi nantly expressed alleles, ALAD1 and ALAD2 (dbSNP ID rs1800435) (Battistuzzi et al. 1981). The difference between these two alleles lies in a single G→C transversion mutation of nucleotide 177 in ALAD2; the allozyme resulting from the ALAD2 allele contains the substitution of a neutral aspar agine for a positively charged lysine at residue 59 (Wetmur et al. 1991). Three differently charged allozymes, ALAD11, 12, and 22, result from the expression of the ALAD1 and ALAD2 alleles, which have different affinities for lead (Bergdahl et al. 1997). The frequen cies of the ALAD1 and ALAD2 alleles in sev eral white populations have been estimated to be 0.9 and 0.1, respectively, whereas Asian and African populations have lower ALAD2 allele frequencies (Kelada et al. 2001). It is well known that individuals vary greatly in susceptibility to lead toxicity, and genetic sus ceptibility has often been cited as the prob able cause for such variation (Kelada et al. 2001). In a review of occupational studies in which lead exposure was relatively high, the rs1800435 polymorphism in the ALAD gene was positively associated with BLL; however, no association has been found between ALAD and BLL among environmentally exposed adults with BLL < 10 µg/dL (Scinicariello et al. 2007).
Few studies have addressed the relation ship between ALAD genotype status and BLL and BP, and the results have been inconclu sive. A study conducted among Korean lead smelter workers (n = 798; mean BLL = 32.0 µg/dL) showed that ALAD2 carriers had a sta tistically significantly increased systolic BP at occupational lead exposure levels compared with ALAD1 homozygous carriers . A previous study conducted among 691 members of a construction trade union (mean BLL = 7.78 µg/dL) found that systolic BP and diastolic BP were increased in ALAD2 carriers compared with members homozygous for the ALAD1 allele. This difference, however, was not statistically significant (Smith et al. 1995).
Currently, the Centers for Disease Control and Prevention (CDC) designates 10 µg/dL as a BLL of concern and has formulated guide lines for environmental and educational inter vention in children at this level (CDC 2002). However, no corresponding CDC guidance exists for BLL measured in adults.
The objective of our study is to determine, in a large, nationally representative sample [Third National Health andNutrition Examination Survey (NHANES III), 1988-1994], whether the relationship between BP and lead exposure differs by ALAD genotype status.

Materials and Methods
NHANES III is a survey representative of the civilian, noninstitutionalized U.S. population and was conducted in 1988-1994 (National Center for Health Statistics 1994). The sample was selected through a complex, multistage probability design and included oversampling of nonHispanic blacks, Mexican Americans, children, and the elderly (National Center for Health Statistics 1994). Sample weights account of differential probability of selec tion and nonresponse and are poststratified volume 118 | number 2 | February 2010 • Environmental Health Perspectives to census population estimates. Individuals participated in an interview conducted at home and also received an extensive physical examination performed at a mobile examina tion center (MEC), which included blood and urine collection. The total number of partici pants from the second phase of NHANES III (1991)(1992)(1993)(1994) was 16,530. The NHANES III DNA bank contains specimens from 7,159 participants ≥ 12 years of age who were exam ined during the second phase of NHANES III (1991)(1992)(1993)(1994) (Chang et al. 2009). The sam ple weights were derived from the NHANES III phase 2 MECexamined sample weights, and they were recalculated using previously described methods (Lohr 1999) for the 7,159 participants for whom DNA was available to avoid nonresponse bias for the NHANES III genetic data.
The present study included participants in the NHANES III DNA bank ≥ 17 years of age and selfreported as nonHispanic white, nonHispanic black, or Mexican American (n = 6,016). Further details regarding the NHANES III DNA bank have been previ ously described (Chang et al. 2009).
Definitions and measurements of variables. Outcome variables. Hypertension status and two BP measures (systolic and diastolic BP) were the outcome measures. Hypertension was defined as selfreported current use of an antihypertensive medication, systolic BP of ≥ 140 mmHg, or diastolic BP of ≥ 90 mmHg. Systolic and diastolic BP levels were examined for persons not currently taking antihyperten sive medication. Hypertension was treated as a dichotomous variable, and the two BP meas ures were treated as continuous variables.
BP for individual participants was cal culated as the average of all available meas urements (maximum = 6) from the home interview and MEC.
BP measurements were performed in the MEC by the physician on children (5-19 years of age) and adults (≥ 20 years of age); the interviewers took measurements on adults only (≥ 17 years of age) in the household. More details can be found at the reference manual of BP quality control program (CDC 1991).
Other variables. BLL was determined using graphite furnace atomic absorption spectrophotometry. Total serum calcium was measured by ionselective electrodes. Serum creatinine was measured by the modi fied kinetic Jaffe reaction using a Hitachi 737 analyzer (Boehringer Mannheim Corp., Indianapolis, IN). Glycosylated hemoglobin (HbA1C) was measured using ion exchange chromatography. Details of the laboratory pro tocols for each of these measures can be found on the CDC National Center for Health Statistics Web site (Gunter et al. 1996).
Phenotypic covariates analyzed in the regression analysis included the continuous variables of body mass index (BMI, calculated as weight in kilograms divided by the square of height in meters) and serum creatinine levels and the categorical variables of age (17-39, 40-59, and ≥ 60 years), sex, education (< high school, completed high school, some college), cigarette smoking history (current, former, or never), and weekly alcohol intake (none, < 4, and ≥ 4 drinks per week).
ALAD genotyping methods. The ALAD polymorphism was genotyped at the Core Genotyping Facility, National Cancer Institute, National Institutes of Health (Bethesda, MD), using TaqMan assays (Applied Biosystems, Foster City, CA). Detailed information on genotyping methods and quality control methods have been previ ously described (Chang et al. 2009). Briefly, the quality of the genetic data was assured through the use of water controls, DNA sam ples with known genotypes, blinded repli cates of approximately 6% of samples, and tests for deviations from Hardy-Weinberg proportions.
Statistical analysis. We used multivariable logistic and linear regression to examine the relationships among hypertension (and BP measures), BLL, and the ALAD polymorphism in nonHispanic white, nonHispanic black, and MexicanAmerican race/ethnicity catego ries. Participants were categorized into BLL quartiles based on the weighted population distribution. The ALAD polymorphism was assessed assuming a dominant model (ALAD1/ ALAD1 vs. ALAD1/ALAD2 and ALAD2/ ALAD2) where the homozygous major allele (ALAD1) was the reference group and the heterozygous plus homozygous for the minor allele (ALAD2) was the evaluated group.
Models were analyzed separately for the three outcomes of interest as indicated above. Effect modification by ALAD in the rela tionship between blood levels and hyperten sion outcome was examined through models stratified by BLL quartiles: BLL = 0.7-1.4, 1.5-2.3. 2.4-3.7. 3.8-52.9 µg/dL. Also, analyses were run excluding from the fourth quartile individuals with BLL > 10.0 µg/dL (BLL range, 3.8-10.0 µg/dL). We used the multivariable linear regres sion to examine the relationships among BP measures (systolic and diastolic BP), ALAD genotype status, logtransformed BLL, and the interaction term between logtransformed BLL and ALAD genotype status. The BLLs were logtransformed (natural logarithms) because the lead levels in blood were right skewed. The change in BP associated with a doubling of the BLL was calculated by multiplying the regres sion coefficient by 0.69 (Nawrot et al. 2002).
We conducted multivariable regression models adjusting for the potential risk fac tors age, sex, education, smoking status, alco hol intake, BMI, serum creatinine levels (as a marker of kidney function), serum calcium, glycosylated hemoglobin (a timeintegrated marker of average glycemia during the previ ous 3 months), and hematocrit. These risk factors have previously shown to be associ ated with BP and hypertension (Burt et al. 1995) and with lead and BP (Nash et al. 2003;Vupputuri et al. 2003), including hematocrit (Hense et al. 1993) and alcohol intake (Hense et al. 1993(Hense et al. , 1994. Statistical analyses were performed using SAS version 9.1 (SAS Institute Inc., Cary, NC) and SAScallable SUDAAN version 9.01 (Research Triangle Institute, Research Triangle Park, NC) to account for the NHANES III complex sample design. pValues from Satterthwaite statistics were presented at the significance level of 0.05. Table 1 illustrates the characteristics of partici pants (n = 6,016) from the NHANES III DNA bank weighted to be representative of the U.S. population. The mean age of the population was 44 years (data not shown). Approximately 52% of individuals were female. NonHispanic whites accounted for 81.6% of the total popu lation. Approximately 43%, 47%, and 49% of the people reported that they had attended some college, never used alcohol, and never smoked, respectively. The mean (± SE) for systolic BP was 119.0 ± 0.39 mmHg and for diastolic BP was 73.5 ± 0.27 mmHg. Hypertension was observed in 22.7% of the population. For those not currently treated for hypertension, 9.2% and 5.5%, had systolic (≥ 140 mmHg) and diastolic (≥ 90 mmHg) hypertension, respectively. We estimated the mean (± SE) BLL to be 2.98 ± 0.09 µg/dL. Table 1 also summarizes the characteristic of the study participants by race/ethnicity. Table 2 summarizes the characteristics of the study participants by BLL quartile. Individuals in the highest BLL quartiles tended to be older, male, current smokers, regular drinkers, less educated, and more likely to have hypertension and to have increased serum creatinine, hematocrit, and systolic and dia stolic levels of BP. Table 3 presents the ALAD genotypes and mean BLL by ALAD genotype and by BLL quartile for each of the three major race/ ethnic groups in the United States. ALAD2 carri ers comprised 13.6% of the total population, 15.6% (370 of 2,017) of nonHispanic whites, 2.6% (49 of 1,621) of nonHispanic blacks, and 8.8% (137 of 1,609) of Mexican Americans. Among participants with the ALAD1-1 genotype, we observed a slightly higher BLL for all race/ethnic subgroups com pared with the participants with ALAD1-2   Table 4 shows adjusted prevalence odds ratio (POR) for hyperten sion for each of the three major race/eth nic subgroups in the U.S. population. We observed no difference between BLL quartile and hypertension in nonHispanic whites and Mexican Americans. However, in the non Hispanic black population, individuals in the second, third, and fourth BLL quartiles had a significant adjusted POR (1.83, 2.38, and 2.93, respectively) for hypertension compared with the lowest quartile. We found no signifi cant associations of ALAD and hypertension in any of the race/ethnic groups. Table 5 presents the adjusted PORs for hypertension comparing ALAD2 car riers and ALAD1 homozygous individuals stratified by BLL quartile and by race/eth nicity. In the nonHispanic white popula tion, ALAD2 carriers in the highest BLL quartile (3.8-52.9 µg/L) had a significantly higher POR [2.00; 95% confidence inter val (CI), 1.12-3.55] of hypertension than did the subjects who were ALAD1 homozy gous. Moreover, when we excluded from the fourth quartile those with BLL > 10.0 µg/dL, we still observed a significant increase in risk of hypertension (POR = 1.86; 95% CI%, 1.00-3.49) (data not shown).

Results
Systolic and diastolic BP. We used multi variate linear regression models to assess asso ciations with systolic and diastolic BP in each race/ethnic group for individuals who were not currently taking medications for high BP (Table 6). We found a significant interac tion between BLL and ALAD in relation to systolic BP in two race/ethnic subgroups (p = 0.04 for nonHispanic whites, and p = 0.02 for nonHispanic blacks). In multivariate regression analyses for diastolic BP, the inter action between ALAD and BLL was not sig nificant in any of the three race/ethnic groups (Table 6).

Discussion
Our results for the association of blood lead with BP and with hypertension in the race/ethnic groups in the U.S. population is consistent with previous NHANES III studies Nash et al. 2003;Vupputuri et al. 2003). Estimates from the coefficient regressions for blood lead (Table  6) predicted that a 2fold increase of BLL was correlated with increases of 1.76 mmHg (95% CI, 1.06-2.47) and 0.72 mmHg (95% CI, 0.19-1.26) in systolic BP for nonHispanic blacks and whites, respectively. Although we did not observe an association with BLL and hypertension in nonHispanic whites and Mexican Americans, we did find an increased prevalence of hypertension in nonHispanic blacks in the second, third, and fourth lead quartile compared with those in the lowest BLL quartile. A previous evaluation of lead levels and hypertension in the NHANES III population by Vupputuri et al. (2003) also showed that black and white women with BLL ≥ 5 µg/dL had statistically significantly higher odds ratio for having hypertension compared with those with BLL < 5 µg/dL.
We found that within the highest quartile of lead for the nonHispanic white population, the prevalence of hypertension was significantly higher among ALAD2 carriers compared with ALAD1 homozygote individuals. In addi tion, estimates from regression coefficients of the interaction terms shown in Table 6 indi cate that, for a doubling of BLL, systolic BP increased by an estimated 2.46 mmHg for ALAD1-2/2-2 individuals and 0.72 mmHg for ALAD1 homozygous individuals for the non Hispanic white population. In contrast, for nonHispanic black individuals, for a doubling of BLL, systolic BP decreased by an estimated 4.04 mmHg for ALAD1-2/2-2 individuals and increased by 1.76 mmHg for ALAD1 homozy gous individuals. This finding may not be reli able for nonHispanic blacks because there were substantially fewer ALAD2 carriers in this population in our study (n < 50).
Of the two previous studies on ALAD, BLL, and BP, one reported an increase in sys tolic BP among Korean ALAD2 carriers who were occupationally exposed to lead , whereas the other reported that ALAD was not associated with systolic BP (Smith et al. 1995).
The mechanisms of leadinduced hyper tension are not well characterized. One hypothesis is that lead induces hypertension through direct effects on the kidney . Another hypothesis is that the accumulation of lead in the walls of arter ies results in arterial stiffness, which induces hypertension. Lead has been reported to both accumulate in the human aorta (Poklis 1975;Schroeder and Tipton 1968) and contribute to the increase in pulse pressure that occurs with aging (Perlstein et al 2007). Elevated aortic stiffness is also known to induce high systolic BP and increase pulse pressure (O'Rourke and Mancia 1999). Finally, it is also possible that lead may alter BP by interference with vascular   signaling pathways. Endothelial nitric oxide (NO) regulates vascular function, and the dis ruption of the NO activity is important in the development of hypertension (Chowdhary and Townend 2001). Lead exposure has been reported to significantly inhibit endothelial NO production , as well as to cause NO inactivation by increasing oxi dative stress, thereby decreasing NO availabil ity (Vaziri and Ding 2001;Vaziri et al. 1999). There was no association between ALAD2 carriers and mean BLL in any of the race/eth nicities. However, in general, ALAD2 carri ers had a lower mean BLL than did ALAD1 homozygous subjects, although these differ ences were not statistically significant. When stratified by BLL quartile, only nonHispanic white ALAD2 carriers had a marginally higher mean BLL than did ALAD1 homozygous sub jects in the fourth BLL quartile. These data are in agreement with the view that ALAD2 allele would significantly affect BLL not at low exposure levels but only at higher levels, when other leadbinding sites have been satu rated (Schwartz et al. 1995;Scinicariello et al. 2007). Montenegro et al. (2006) reported that, although ALAD2 carriers had no significantly lower mean BLL than did ALAD1 homozy gous individuals, ALAD2 carriers had a sig nificantly higher plasma lead level compared with homozygous ALAD1 subjects. Therefore, it may be possible that nonHispanic white ALAD2 carriers have higher levels of plasma lead compared with ALAD1 homozygote indi viduals. Consequently, the higher plasma lead, interacting with other molecular BP regulatory systems, may be responsible for the observed increases in systolic BP in ALAD2 carriers.
The present study has several limitations. The exclusion from our study of persons who reported taking medication for hypertension may have diluted the strength and magnitude of associations in our analysis for systolic and diastolic BP. Second, although we controlled for many of the known factors associated with BP and hypertension, other variables such as serum selenium (Telisman et al. 2001), serum zinc (Schwartz 1991), and blood cadmium (NavasAcien et al. 2004) might have influ enced our findings. Blood cadmium was not measured by NHANES III. However, urinary cadmium, a measured variable in NHANES III, was not a significant variable for hyperten sion or for systolic or diastolic BP (data not shown). Urinary cadmium reflects cadmium concentration in the renal cortex and is a bio marker of both ongoing and longterm cad mium exposure, whereas blood cadmium is a biomarker of ongoing exposure (Agency for Toxic Substances and Disease Registry 1999). Therefore, it seems unlikely that including blood cadmium in our models would have changed our finding. Our results are based on BLL; because approximately 95% of the total body burden of lead is present in the skeleton, a preferred measure of chronic body burden would be bone lead (Hu et al. 1996). However, the measurement of bone lead in a large sample size, such as NHANES, is not feasible, and blood lead is known to be associated with bone lead (Gwiazda et al. 2005;Todd et al. 2001).
Currently, the CDC designates 10 µg/dL as a BLL of concern and has formulated guide lines for environmental and educational inter vention in children at this level (CDC 2002). No corresponding CDC guidance exists for BLL measured in adults. However, in the past few decades the presence of lead in the envi ronment has steadily declined. In the adult U.S. population, mean BLL measured in NHANES surveys conducted in 1976NHANES surveys conducted in -1980NHANES surveys conducted in , 1988NHANES surveys conducted in -1991NHANES surveys conducted in , and 1999NHANES surveys conducted in -2002 decreased from 13.1 µg/dL to 3.0 µg/dL and to 01.64 µg/dL, respectively (Muntner et al. 2005;Pirkle et al 1994). This positive and welcome decline has steadily continued with a geometric mean BLL of 1.41 µg/dL in the U.S. adult popu lation ≥ 20 years of age as measured in the NHANES survey conducted in .

Conclusions
We examined the modification by ALAD in the association of BLL and BP and hyperten sion in the U.S. population. We found that within the highest quartile of lead for the non Hispanic white population, the prevalence of hypertension was significantly higher among ALAD2 carriers compared with ALAD1 homozygotes. In addition, we found that ALAD2 carriers in nonHispanic whites may experience a more pronounced effect of lead on systolic BP than do homozygous ALAD1 carriers. These results underscore the impor tance of reducing environmental sources of lead exposure in the U.S. population, and this should remain a major public health prior ity based on consistent evidence of increased health risks (Kosnett et al. 2007;Schwartz and Hu 2007). Given the high frequency of ALAD2 carriers (15.6%) in the nonHispanic white population, and the crosssectional nature of the present study, prospective studies are needed to confirm and elucidate the role of ALAD polymorphism and these associations.