Prenatal Lead Exposure, Type 2 Diabetes, and Cardiometabolic Risk Factors in Mexican Children at Age 10–18 Years

Abstract

Context

Several cross-sectional studies have assessed the association of lead exposure with type 2 diabetes and cardiometabolic risk factors in adults; however, studies of such associations in childhood are rare.

Objective

We assessed the prospective associations of prenatal exposure to lead with type 2 diabetes and cardiometabolic risk factors in children.

Design

The Early Life Exposure in Mexico to Environmental Toxicants is a birth cohort study of pregnant women and their offspring.

Setting

Public hospitals in Mexico City.

Patients or Other Participants

Women were recruited during pregnancy; their offspring were recruited for a follow-up visit at age 10 to 18 years (n = 369).

Main Outcome Measures

We measured fasting serum markers of type 2 diabetes and cardiometabolic risk factors in children, including fasting glucose, insulin, and lipids. The index of insulin resistance was calculated.

Results

The geometric mean of maternal blood lead levels (BLLs) during pregnancy was 4.3 µg/dL (95% confidence interval [CI]): 4.0-4.6 µg/dL) in the entire sample. In boys, those with maternal BLLs ≥ 5 µg/dL (compared with those with BLLs < 5 µg/dL) had significantly lower z scores for total cholesterol (β = -0.41, 95% CI: -0.71, -0.12), high-density lipoprotein cholesterol (β = -0.32, 95% CI: -0.59, -0.05), and low-density lipoprotein cholesterol (β = -0.52, 95% CI: -0.81, -0.22), adjusting for covariates. No associations were detected in girls.

Conclusions

In our study, we found that higher prenatal exposure to lead was associated with lower levels of cholesterol in children following a sex-specific pattern. Further studies with a larger sample size that examine whether sex is a potential modifier are needed to confirm our findings.

The prevalence of type 2 diabetes and metabolic syndrome in adults has been growing steadily worldwide with the highest rising rates seen in low- and moderate-income populations in the past decade (1, 2). Previously considered metabolic disorders exclusively for middle-aged and older adults, type 2 diabetes and metabolic syndrome have been seen more frequently in children, adolescents, and young adults (3–5). Type 2 diabetes and metabolic syndrome in childhood have been related to increased risk of morbidity and mortality in adulthood (6, 7). In 2015, it was reported by the International Diabetes Federation that Mexico is 1 of the countries that have the highest prevalence of diabetes; 11.4 million Mexicans have been diagnosed with this disease (8). Using the Mexico National Health and Nutrition Survey, researchers have estimated that the lifetime risk of pediatric diabetes for Mexican children born in 2010 would be 34.3% to 53.3% (9). In addition, current evidence suggests that Mexicans and Mexican Americans with diabetes have increased mortality risk compared with non-Hispanic whites, which further increases the burden of this disease for this group (10). Depending on the definition of the metabolic syndrome, it occurs in about 2.4% to 45.9% children in Mexico (11), which shows that the existing diagnostic criteria are not effective in determining childhood metabolic syndrome among this population.

The increase in the incidence of type 2 diabetes and metabolic syndrome is widely attributed to changes in diet, a sedentary lifestyle, and genetic predisposition; however, these factors do not fully account for the rapid outbreak of type 2 diabetes and metabolic syndrome. Increasing evidence has shown that environmental exposure to heavy metals, such as lead, contributes to the rising in the prevalence of type 2 diabetes and metabolic syndrome (12–14), and children are particularly susceptible to its toxicity (15). The mechanisms by which lead affects the development of type 2 diabetes and metabolic syndrome are not clear: however, lead may induce oxidative stress (16), which plays an essential role in the pathophysiology of obesity, type 2 diabetes, and dyslipidemia (17). Although the environmental lead concentrations in Mexico has been reduced since the elimination of leaded gasoline in the 1990s, lead contamination continues to be a major public health issue because of its persistence in the environment and the extensive use of lead-glazed ceramic cookware (18).

Several studies have assessed the association between lead exposure and type 2 diabetes and cardiometabolic risk factors; however, their findings were inconsistent. Specifically, previous human studies have shown that increasing blood lead concentrations were significantly associated with increased levels of triglycerides (19, 20), total cholesterol (20) and fasting glucose (20, 21), and decreased prevalence of low high-density lipoprotein cholesterol (HDL-C) (HDL-C <50 mg/dL for females or <40 mg/dL for males) (22). In contrast, other epidemiological studies reported no associations with these individual risk factors or with the prevalence of diabetes and metabolic syndrome (23–26).

In summary, the majority of published studies were cross-sectional (19, 21–24, 26), with only 1 prospective study that included prenatal exposure to lead using maternal blood lead concentrations during pregnancy (25). In addition, existing studies have focused on characterizing the association between lead exposure and type 2 diabetes and cardiometabolic risk factors in adults (19, 21–24, 26), with 2 studies that examined such associations in children during early childhood (25) and adolescence (20). Given that previous analyses have reported sex differences in the regulation of glucose, insulin sensitivity, and lipid levels (27, 28), it is reasonable to assess whether sex modifies the association between exposure to lead and type 2 diabetes and cardiometabolic risk factors. To address these aforementioned knowledge gaps, in the present analysis, we aimed at examining the association of prenatal exposure to lead during pregnancy as measured by maternal blood lead concentrations with multiple type 2 diabetes and cardiometabolic risk factors in Mexican boys and girls at age 10 to 18 years.

Materials and Methods

Study populations

We used data collected from participants of the Early Life Exposures in Mexico to ENvironmental Toxicants (ELEMENT) study, which consists of 3 sequentially enrolled birth cohorts of pregnant women (29). Briefly, participants of cohorts 1 and 2B did not have archived biological samples and were excluded. Participants of cohorts 2A and 3 were recruited from 2 public hospitals (Mexican Social Security Institute and the National Institute of Perinatology) in Mexico City that serve low-to-moderate income populations. Cohort 2A was an observational study that recruited participants between 1997 and 1999. Cohort 3 was a randomized clinical trial of calcium supplementation starting from the second trimester that recruited participants between 2001 and 2003. Detailed information regarding the recruitment, eligibility criteria, and collection of maternal information of the ELEMENT study have been previously described (30, 31). Starting in 2015, a subset of the offspring (n = 400) from cohorts 2A and 3 were re-recruited if they were in peripubertal period (age 10–18 years) and were measured for serum makers of type 2 diabetes and cardiometabolic risk factors. During the follow-up visit, children completed anthropometric assessments and interview-based questionnaires. Of the 400 participants, 369 who had measured prenatal blood lead concentrations during the first trimester were included in the final analyses.

Research protocols of this study were approved by the institutional review board at the University of Michigan and the Mexico National Institute of Public Health. We obtained informed consent from mothers and informed assent from children before enrollment.

Type 2 diabetes and cardiometabolic risk factors

Fasting glucose, insulin, and lipids were measured in serum at the Michigan Diabetes Research Center Chemistry Laboratory. Specifically, fasting glucose was assessed using an enzymatic method (Sekisui Diagnostics, LLC, Lexington, MA). The concentrations of insulin were assessed by an immunoturbidimetric assay (Sekisui Diagnostics, LLC). Triglycerides and total cholesterol were quantified via enzymatic colorimetric method on a Cobas Mira automated chemistry analyzer (Roche Diagnostics, Indianapolis, IN). The levels of HDL-C and low-density lipoprotein cholesterol (LDL-C) were obtained by using direct HDL-C (Roche Diagnostics) and direct LDL-C assays (Equal Diagnostics, Exton, PA), respectively. In addition, the index of insulin resistance (i.e., homeostatic model assessment of insulin resistance [HOMA-IR]) was calculated from fasting glucose and insulin using the following formula: HOMA-IR = insulin (µU/mL) × glucose (mg/dL)/405 (32). All these assays were in agreement with the National Cholesterol Education Program guidelines. All the serum markers were above the limit of detection (LOD).

Prenatal blood lead

A sample of whole blood was collected from each pregnant woman during the first trimester visit by trained staff using standardized protocols. All the samples were stored in trace-metal–free tubes. The concentrations of lead in whole blood were measured at the Trace Metal Laboratory at the American British Cowdray Hospital in Mexico City using graphite-furnace atomic-absorption spectroscopy (Model 3000; Perkin-Elmer, Chelmsford, MA) according to previously described techniques (33). The LOD was 0.1 μg/dL. In this study population, 7 blood samples that had values below LOD were input as 0.1 μg/dL.

Statistical analysis

Univariate and bivariate analyses were conducted. Levels of blood lead, HOMA-IR, and triglyceride were right skewed. We categorized the prenatal blood lead levels (BLLs) into ≥5 μg/dL and <5 µg/dL according to the Centers for Disease Control and Prevention’s and the Ministry of Health of Mexico’s reference level of blood lead at 5 µg/dL for public health interventions (15, 34). It is important to demonstrate the clinical significance of the reference level by directly linking it to child health outcomes. We did not categorize lead levels by quartile because it would reduce our power to detect significant associations. HOMA-IR and triglyceride were log-transformed to approximately conform to normality. Mean ± standard deviations (SD) or proportions (%) were calculated for variables with normal distribution. Geometric mean and 95% confidence intervals (CI) were calculated for variables with skewed distribution. Student t test or χ ² test was used to determine the significance of mean differences in the demographic characteristics by the 2 groups (prenatal BLLs ≥5 µg/dL vs <5 µg/dL). All type 2 diabetes and cardiometabolic risk factors were converted to standardized age- and sex-specific z scores before the final analyses.

Multivariable linear regression models were fitted to examine the association between maternal BLLs (≥5 µg/dL vs <5 µg/dL) during pregnancy and z scores for fasting glucose, HOMA-IR, triglycerides, total cholesterol, HDL-C, or LDL-C. We chose covariates a priori based on biological relevance as known predictors of type 2 diabetes and cardiometabolic risk factors or potential confounders. All models were adjusted for child age, sex, body mass index (BMI) z score, and number of siblings at birth, maternal age, marital status, education (a proxy for socioeconomic status), and smoking history. Diet is an important predictor of type 2 diabetes and cardiometabolic risk; however, we did not consider diet as a covariate in this study because it may mediate the associations by lowering the food consumption in response to lead exposure (35).

The potential modifying effect of children’s sex on the association between prenatal BLLs and type 2 diabetes and cardiometabolic risk factors was evaluated by including interaction terms between lead and sex for each outcome. Because we found some evidence of children’s sex as a modifier (P for sex interaction ≤0.10), we also examined the sex-specific associations of prenatal blood concentrations with each risk factor. We defined the statistical significance by an alpha level ≤0.10 for interaction terms and ≤0.05 for all other estimates. All statistical analyses were performed using SAS (version 9.4; SAS Institute Inc., Cary, NC).

Sensitivity analyses

In sensitivity analyses, we reanalyzed all the models with additional adjustment for pubertal stage, which has been associated with alterations in the metabolic profiles (36). An experienced pediatrician evaluated the stage of pubertal development in children using Tanner stage ranging from 1 to 5 for pubic hair growth for both boys and girls (37). In addition, we repeated our analyses with further adjustment for concurrent childhood BLLs. Finally, we reanalyzed all the models with further adjustment for maternal diet and alcohol consumption during pregnancy.

Results

The final analyses included 369 children with a total of 186 girls and 183 boys at ages 10 to 18 years. In the total sample, the mean age for children was 13.7 years (SD, 1.9) (Table 1). On average, mothers were 26.7 (SD, 5.6) years old at delivery, had 11.0 (SD, 2.9) years of education, 72.1% were married, and 50.3% had a history of smoking. Child age was significantly older for those whose maternal BLLs ≥5 µg/dL versus <5 µg/dL (14.4 vs 13.1 years, P < 0.0001). No significant differences in other demographic characteristics across categories of maternal BLLs were detected. The geometric mean (95% CI) of maternal BLLs during pregnancy in boys and girls combined was 4.3 µg/dL (4.0-4.6) (Table 2). In this study sample, 17.1% of children were at stage 1 (prepubertal), 25.8% at stage 2 (pubertal onset), 25.0% at stage 3, 20.5% at stage 4, and 11.5% were at stage 5 (adult).

Table 1.

Demographic Characteristics of Participants According to Prenatal Blood Lead Levels in Mexico City^a

		Prenatal Blood Lead Level
	Overall (n = 369)	≥5 µg/dL (n = 171)	<5 µg/dL (n = 198)
	Mean (SD) or N (%)	Mean (SD) or N (%)	Mean (SD) or N (%)
Child characteristics
Age, yb	13.7 (1.9)	14.4 (2.0)	13.1 (1.6)
Female	186 (50.4%)	87 (50.9%)	99 (50.0%)
BMI z score	0.6 (1.2)	0.5 (1.3)	0.6 (1.2)
No. siblings at birth	2.0 (1.0)	2.0 (1.0)	2.0 (1.0)
Maternal characteristics
Maternal age, y	26.7 (5.6)	26.9 (5.9)	26.4 (5.3)
Maternal education, y	11.0 (2.9)	10.9 (2.9)	11.0 (3.0)
Marital status
Yes	266 (72.1%)	130 (76.0%)	136 (68.7%)
No	103 (27.9%)	41 (24.0%)	62 (31.3%)
Smoking history
Ever	181 (50.3%)	84 (50.9%)	97 (49.7%)
Never	179 (49.7%)	81 (49.1%)	98 (50.3%)

Abbreviations: BMI, body mass index; SD, standard deviation.

^aMissing information: maternal smoking history (n = 9).

^bStudent t test with P value ≤ 0.05.

Table 2.

Distributions of T2DM and Cardiometabolic Risk Factors and Prenatal Blood Lead Levels Overall and by Child Sex

	Overall (n = 369)	Boys (n = 183)	Girls (n = 186)
	Mean (SD)	Mean (SD)	Mean (SD)
Risk factors
Fasting glucose, mg/dL	77.8 (7.2)	79.0 (7.2)	76.7 (6.9)
HOMA-IRa	3.2 (3.0-3.4)	3.0 (2.8-3.3)	3.4 (3.2-3.7)
Triglycerides, mg/dLa	93.1 (88.8-97.6)	87.5 (81.7-93.7)	99.0 (92.8-105.6)
Total cholesterol, mg/dL	155.5 (26.3)	150.5 (26.6)	160.3 (25.2)
HDL-C, mg/dL	43.0 (8.7)	42.0 (7.8)	44.0 (9.3)
LDL-C, mg/dL	91.6 (20.8)	88.9 (21.2)	94.2 (20.1)
Lead biomarker
Prenatal blood lead, µg/dLa	4.3 (4.0-4.6)	4.2 (3.7-4.7)	4.4 (3.9-4.9)

Abbreviations: HDL, high-density lipoprotein cholesterol; HOMA-IR, homeostatic model assessment of insulin resistance; LDL-C, low-density lipoprotein cholesterol; T2DM, type 2 diabetes.

^aGeometric mean ± 95% confidence interval.

Table 3 shows the adjusted associations between prenatal lead exposure during pregnancy and type 2 diabetes and cardiometabolic risk factors in the total sample. Children with higher maternal BLLs (≥5 µg/dL) had a significantly lower total cholesterol and LDL-C z scores than those with lower maternal BLLs (<5 µg/dL), adjusted for child age, sex, BMI z score, and number of siblings at birth, maternal age, marital status, education, and smoking history. We also observed a borderline significant association of higher maternal BLLs with lower levels of HDL-C (P = 0.05).

Table 3.

Adjusted Overall Associations Between Prenatal Lead Exposure (≥5 µg/dL vs <5 µg/dL) and T2DM and Cardiometabolic Risk Factors in Children at Age 10 to 18 Years^a

	Overall
	β (95% CI)	P	P for Sex Interaction
Fasting glucose z score
≥5 µg/dL	–0.05 (–0.69, 0.60)	0.89	0.96
HOMA-IR z score
≥5 µg/dL	–0.11 (–0.63, 0.42)	0.69	0.69
Triglycerides z score
≥5 µg/dL	0.58 (–0.05, 1.20)	0.07	0.14
Total cholesterol z score
≥5 µg/dL	–0.76 (–1.38, –0.13)	0.02	0.10
HDL-C z score
≥5 µg/dL	–0.64 (–1.28, 0.01)	0.05	0.10
LDL-C z score
≥5 µg/dL	–0.96 (–1.59, –0.33)	0.003	0.04

Abbreviations: CI, confidence interval; HDL, high-density lipoprotein cholesterol; HOMA-IR, homeostatic model assessment of insulin resistance; LDL-C, low-density lipoprotein cholesterol; T2DM, type 2 diabetes.

^aAdjusted for child age, sex, body mass index z score, and number of siblings at birth, maternal age, marital status, education, and smoking history.

As seen in Table 4, when stratifying by children’s sex, multiple linear regression models showed that boys with maternal BLLs ≥5 µg/dL had significantly lower z scores for total cholesterol (β = -0.41; 95% CI, -0.71, -0.12; P for sex interaction = 0.10), HDL-C (β = -0.32; 95% CI, -0.59, -0.05; P for sex interaction = 0.10), and LDL-C (β = -0.52; 95% CI, -0.81, -0.22; P for sex interaction = 0.04) than those with maternal BLLs < 5 µg/dL. However, maternal BLLs were not significantly associated with fasting glucose, HOMA-IR, or triglycerides in boys. No associations were found in girls.

Table 4.

Adjusted Sex-Specific Associations Between Prenatal Lead Exposure and T2DM and Cardiometabolic Risk Factors in Children at Age 10 to 18 Years^a

	Boys		Girls
	≥5 µg/dL (N = 81) vs <5 µg/dL (N = 98)		≥5 µg/dL (N = 84) vs <5 µg/dL (N = 97)
	β (95% CI)	P	β (95% CI)	P
Fasting glucose z score
≥5 µg/dL	–0.05 (–0.34, 0.25)	0.76	–0.06 (–0.35, 0.23)	0.69
HOMA-IR z score
≥5 µg/dL	–0.04 (–0.28, 0.20)	0.74	0.04 (–0.19, 0.27)	0.73
Triglycerides z score
≥5 µg/dL	0.26 (–0.04, 0.55)	0.09	–0.03 (–0.30, 0.24)	0.81
Total cholesterol z score
≥5 µg/dL	–0.41 (–0.71, –0.12)	0.007	–0.11 (–0.38, 0.16)	0.40
HDL-C z score
≥5 µg/dL	–0.32 (–0.59, –0.05)	0.02	0.01 (–0.30, 0.32)	0.95
LDL-C z score
≥5 µg/dL	–0.52 (–0.81, –0.22)	0.007	–0.15 (–0.43, 0.12)	0.27

Abbreviations: CI, confidence interval; HDL, high-density lipoprotein cholesterol; HOMA-IR, homeostatic model assessment of insulin resistance; LDL-C, low-density lipoprotein cholesterol; T2DM, type 2 diabetes.

^aMissing information: maternal smoking history (n = 9); adjusted for child age, body mass index z score, and number of siblings at birth, maternal age, marital status, education, and smoking history.

In sensitivity analyses, our results were not significantly changed when including concurrent childhood BLLs or maternal alcohol consumption during pregnancy. When further adjusting for pubertal stage, we found that the association of maternal BLLs with HDL-C z score became borderline significant (P = 0.06) in boys. However, the magnitude, direction, and significance of the associations with other risk factors were not altered appreciably. Similarly, additional adjustment for maternal dietary inflammation index did not significantly change our results; however, the association of maternal BLLs with HDL-C z score became borderline significant (P = 0.08) in boys. All the results from sensitivity analyses are presented in an online repository (38).

Discussion

In our study of 369 mother-offspring pairs living in Mexico City, we investigated the association between prenatal exposure to lead during pregnancy and multiple type 2 diabetes and cardiometabolic risk factors. We found that higher maternal BLLs in the first trimester was significantly associated with lower total cholesterol, HDL-C, and LDL-C concentrations in boys at age 10 to 18 years. However, no associations were observed among girls.

Our findings of negative associations with the levels of total cholesterol, HDL-C, and LDL-C concentrations are somewhat inconsistent with prior epidemiological analyses; previous studies did not examine the potential sex-specific associations. Specifically, Bulka et al. found that higher BLLs (>1.64 vs <0.71 μg/dL) were associated with decreased prevalence of low HDL-C (HDL-C <50 mg/dL for female or <40 mg/dL for male) in 1088 US adults (22). An investigation of 320 Iranian children at ages 10 to 18 years demonstrated that children with higher BLLs (≥0.110 vs ≤0.089 μg/L) had higher total cholesterol (20). In contrast, cross-sectional studies revealed that BLLs were not associated with the levels of HDL-C and/or LDL-C in 150 adults of African descent (BLLs >1.66 μg/dL) (21) and 1405 South Korean adults (geometric mean [95% CI] of BLLs: 2.45 [2.33–2.58] μg/dL) (19). Likewise, Kupsco et al. conducted a prospective study of 411 mother-offspring pairs in Mexico and found no association of maternal BLLs (mean ± SD, 3.7 ± 2.7 μg/dL) in the second trimester with non–HDL-C levels in children aged 4 to 6 years (25). The differences in the study design, age and ethnicity of participants, cutoffs used for exposure levels, and statistical analyses may contribute to these inconsistencies. For example, the majority of the published studies used a cross-sectional design to examine the associations between lead and lipids in adult populations (19, 21, 22). Additionally, the African study (21) was limited by the small sample size and the inclusion of participants from 5 countries. The study by Rhee et al. (19) did not adjust for adiposity, which has been shown to be associated with type 2 diabetes and dyslipidemia (39, 40). For the 2 studies of children or adolescents, the Iranian study (20) was not able to account for the potential confounding effect of socioeconomic status, which is an independent predictor of type 2 diabetes and metabolic syndrome (41, 42). In addition, this analysis (20) used a case-control study design that recruited children with and without metabolic syndrome, which may contribute to the inconsistent findings. Finally, the lipid concentrations measured in Kupsco et al.’s study (25) were from nonfasting samples in the second trimester, which may affect their results, and they included children at a much younger age.

In the present analysis, maternal BLLs were not significantly associated with the levels of glucose, HOMA-IR, or triglycerides. Similar to our results, 1 epidemiological study found that BLLs (mean ± SD, 2.41 ± 1.52 μg/dL) were not associated with HOMA-IR in 1588 adult men and 1596 adult women living in South Korea (23). Cross-sectional studies revealed that lead concentrations were not associated with the levels of glucose in 1274 Canadian pregnant women (>18 years) (BLLs >0.9 vs <0.5 μg/dL) (24) and 2242 Chinese adults (urinary lead levels >4.52 vs <2.12 μg/L) (26). No associations between BLLs and triglycerides concentrations were reported by Kupsco et al. (25) and Bulka et al. (22). On the contrary, higher BLLs (≥0.110 vs ≤0.089 μg/L) were associated with higher levels of triglycerides and fasting glucose in 320 Iranian adolescents (20). The positive associations with the concentrations of fasting glucose and triglycerides were also observed among adults of African descent (21) and South Korean adults (19), respectively.

According to our results, pregnancy is a potentially sensitive period of lead’s impact on cholesterols. This is biologically plausible because lead can pass through the placenta of mothers and store in the developing organs of fetus (43). In addition, the mobilization of lead from maternal stores to fetal circulation increases during pregnancy (43). Furthermore, it has been shown that early gestation is more sensitive to the development of abnormal lipid profiles later in life (44).

Our finding of a negative association between early gestational exposure to lead and HDL-C in boys is possible because lead can induce oxidative stress, which is associated with decreased concentrations of HDL-C (45). Although the biological mechanisms by which lead causes oxidative stress are not clear, it is possible that lead can stimulate oxidative stress via the direct formation of reactive oxygen species and the reduction of the cellular antioxidant defense system by depleting glutathione (16). Furthermore, it has been shown that females may be less sensitive to oxidative stress than males, which could be in part due to the antioxidant properties of estrogen (46). Although the role of oxidative stress in the levels of LDL-C has not been well established, lead may decrease the levels of LDL-C by reducing adiposity (47); in fact, it has been shown that percentage of body fat is significantly associated with the incident hyper-LDL cholesterolemia in men but not in women (48), which could explain the sex-specific associations observed in our study. Another possibility is prenatal exposure to lead reprograms the regulation of cholesterol metabolism (44), which could restrict the ability to synthesize cholesterols later in life. Further studies that examine the modifying effect of children’s sex are needed.

Although the role of dyslipidemia in cardiovascular diseases is widely recognized, the associations of cholesterols with type 2 diabetes are not well characterized (49). Previous studies have shown that both LDL-C and HDL-C are inversely associated with type 2 diabetes (49–51). Therefore, according to our findings, it is reasonable to suggest that lead exposure during early pregnancy may be associated with increased risk of type 2 diabetes in children. Independent of type 2 diabetes, low HDL-C has also been shown to predict the incidence of the myocardial infarction, ischemic stroke, and cardiovascular morbidity (52), whereas high LDL-C may cause atherosclerotic cardiovascular disease, and LDL-C reduction could improve coronary artery disease mortality (53).

Our study has some limitations. First, we were limited by the relatively small sample size, which may have reduced the power to detect significant results. Second, given that our analysis included participants with low-to-moderate income in a Mexican population, these findings may not be generalizable to other populations with different socioeconomic levels or ethnicity. Finally, we were not able to examine the modifying effects of oxidative stress on the observed associations. Despite these limitations, our study has several strengths. We observed sex-specific associations between lead exposure and lipids profiles, which have not been reported by earlier studies. Another strength is the prospective study design that allowed us to establish temporal link. Finally, several potential confounders and covariates associated with exposures or outcomes were considered in the present analysis.

In conclusion, we found that higher prenatal exposure to lead was associated with a decrease in the levels of HDL-C, total cholesterol, and LDL-C among Mexican boys, but not girls. Future studies with a large sample size evaluating the potential modifying effect of sex are needed to confirm our findings.

Glossary

Abbreviations

BLL

blood lead level

BMI

body mass index

CI

confidence interval

HDL-C

high-density lipoprotein cholesterol

HOMA-IR

homeostatic model assessment of insulin resistance

LDL-C

low-density lipoprotein cholesterol

LOD

limit of detection

SD

standard deviation

Additional Information

Disclosure Summary: The authors have no conflicts of interest to disclose.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.