Prenatal blood lead levels and reduced preadolescent glomerular filtration rate: Modification by body mass index

Abstract

Background:

For the developing kidney, the prenatal period may represent a critical window of vulnerability to environmental insults resulting in permanent nephron loss. Given that the majority of nephron formation is complete in the 3^rd trimester, we set out to test whether 1) prenatal lead exposure is associated with decreased preadolescent kidney function and 2) whether preadolescent obesity acts synergistically with early life lead exposure to reduce kidney function.

Methods:

Our study included 453 mother-child pairs participating in the PROGRESS birth cohort. We assessed prenatal blood lead levels (BLLs) in samples collected in the 2^nd and 3^rd trimesters and at delivery, as well as tibial and patellar bone lead measures assessed one-month postpartum. Preadolescent estimated glomerular filtration rate (eGFR) was derived from serum levels of creatinine and/or cystatin C measured at age 8–12 years. We applied linear regression to assess the relationship between prenatal bone and BLL with preadolescent eGFR, and adjusted for covariates including age, sex, BMI z-score, indoor tobacco smoke exposure, and socioeconomic status. We also examined sex-specific associations and tested for effect modification by BMI status.

Results:

We observed null associations between prenatal lead exposure and eGFR. However, in interaction analyses we found that among overweight children, there was an inverse association between BLL (assessed at 2^nd and 3^rd trimester and at delivery) and preadolescent eGFR. For example, among overweight participants, a one ln-unit increase in 2^nd trimester BLL was associated with a 10.5 unit decrease in cystatin C-based eGFR (95% CI: −18.1, −2.8; p = 0.008). Regardless of lead exposure, we also observed null relationships between BMI z-score and eGFR overall, as well as among overweight participants. However, among participants with preadolescent obesity, we observed a significant 5.9-unit decrease in eGFR_. We observed no evidence of sex-specific effects.

Conclusions:

Our findings, if confirmed in other studies, suggest a complex interplay between the combined adverse effects of adiposity and perinatal lead exposure as they relate to adolescent kidney function. Future studies will assess kidney function and adiposity trajectories through adolescence to better understand environmental risk factors for kidney function decline.

Background:

Decreased kidney function is a driver of hypertension, which is a major risk factor for cardiovascular disease, the leading cause of death in the developed world. Furthermore, in the last 30 years, the average weight of a U.S. child has risen by >5 kg.¹ Obesity is a well-known risk factor for hypertension and chronic kidney disease (CKD), and to our knowledge, the convergence of nephrotoxic environmental chemicals and child obesity has not been previously studied. Nephron number at birth varies across individuals² and nephrogenesis is largely complete at 36 weeks’ gestation.³ Beyond 36 weeks, the kidneys continue to grow, but this is typically due to the expansion and maturation of existing nephrons rather than the formation of new nephrons. Low nephron number is an established risk factor for CKD^4,5 and it has been suggested that the prenatal milieu may permanently alter adult organ function.⁵

Lead is an established tubular and glomerular nephrotoxicant.^6–8 Lead can also cross the placenta⁹, and thus maternal exposure results in fetal exposure in utero.¹⁰ Prenatal lead exposure is widespread in the U.S. and Mexico. In Mexico, more than 20% of Mexican mothers have blood lead levels (BLLs) greater than 5 μg/dL during pregnancy which is the CDC guideline level of concern.^11,12 In the systemic circulation, lead has an approximately 28-day half-life, which refers to transfer out of the blood.¹³ While the blood levels are a small percentage of the total body burden, the bloodstream is responsible for distributing lead to other tissues.

In contrast, bone lead levels have a half-life of approximately 25–30 years¹⁴ and can be measured noninvasively by X-ray fluorescence. Bone lead levels are therefore more reflective of cumulative exposure to lead due to the longer half-life.¹⁵ In the context of nephron development, BLLs measured in pregnancy may be particularly informative, as BLLs peak in the late 2^nd and 3^rd trimester secondary to physiologic mobilization of bone calcium stores to meet fetal requirements. Because lead can replace calcium in the bony matrix, the same physiologic processes will remobilize bone lead resulting in increased BLLs later in pregnancy.¹⁶ Given that the majority of nephron formation is completed in the 3^rd trimester, the concomitant increase in maternal BLL coupled with active kidney development may represent a critical window of susceptibility for the developing kidney.^17,18

In addition to lead, overweight and obesity are independent risk factors for kidney dysfunction. Increased adiposity is one of the strongest risk factors for CKD and preclinical kidney disease.^19,20 Obesity leads to kidney injury through glomerular hyperperfusion and hyperfiltration, which can lead to further kidney decline due to insulin resistance, inflammation, and hypertension.²¹ The major increase in childhood obesity in recent decades further underscores the potential importance of the early life environment on kidney function in adulthood.^22,23 Furthermore, our group has shown that lead may act synergistically with established risk factors (such as preterm birth), to negatively influence kidney outcomes.¹²

In this study, we set out to assess whether 1) prenatal lead exposure (assessed in maternal blood and bone matrices) is associated with decreased preadolescent kidney function and 2) whether preadolescent obesity acts synergistically with early life lead exposure to reduce kidney function.

Methods:

PROGRESS Cohort Description:

We conducted an analysis of 453 mother-child pairs participating in the Programming Research in Obesity, Growth, Environment and Social Stressors (PROGRESS) study. PROGRESS is a prospective, ongoing, NIH-funded birth cohort study in Mexico City founded in 2007. Pregnant women receiving care through the Mexican Social Security System (IMSS) were enrolled between July 2007 and February 2011. All participants provided informed and written consent. The study protocols were approved by the institutional review boards of the Brigham and Women’s Hospital, the Icahn School of Medicine at Mount Sinai and the National Institute of Public Health in Mexico.

Women were eligible for enrollment if they were 18 years or older, pregnant at fewer than 20 weeks’ gestation, had access to a telephone, and planned to reside in Mexico City for the following three years. Women were excluded if they had heart or kidney disease, used steroids or anti-epilepsy drugs, or consumed alcohol on a daily basis. All measures of interest were gathered during visits at the 2^nd and 3^rd trimester, delivery, one-month postpartum, and 8–12 years of age.

Of 948 mother-infant pairs assessed, a total of 581 preadolescents attended the 8–12 year-old visit. The study participants for this analysis were restricted to the 453 dyads who had complete measures of serum cystatin C, 2^nd trimester blood lead concentration, and covariates of interest.

Blood Lead Measurements:

Maternal blood was collected at the 2^nd and 3^rd trimester visit. Cord blood lead was also collected after parturition. All blood specimens were drawn in trace metal free tubes and refrigerated at 2–6°C until analysis. Lead concentration was measured by external calibration using the Agilent 8800 ICP Triple Quad (Agilent, Santa Clara, CA) in MS/MS mode in the trace metals laboratory at the Icahn School of Medicine at Mount Sinai. The limit of detection was <0.2μg/dL and the instrument precision (given as %RSD) was approximately 5%. Blinded quality control samples obtained from the Maternal and Child Health Bureau and the Wisconsin State Laboratory of Hygiene Cooperative Blood Lead Proficiency Testing Program showed good precision and accuracy.

Bone Lead Measurements:

One month postpartum, tibia (cortical bone) and patella (trabecular bone) lead concentrations were measured in mothers using a K-shell X-ray fluorescence instrument.²⁴ Lead concentrations were estimated for 30 minutes for each leg, and the measures were averaged by the inverse of the proportion of the measurement error corresponding to each determination. With this instrument, negative values can be obtained when the true bone lead concentration value is close to 0: the instrument produces a continuous unbiased point estimate that fluctuates around the true bone lead value.²⁵ Téllez-Rojo et al. found similar results²⁶ between measurements that included the negative values and those that used simulated estimates randomly generated from a normal distribution. In this analysis, bone lead measurements were treated as a continuous variable in linear regression models.

Cystatin C and Creatinine Measurements:

Cystatin C (previously stored at −70 C) was measured in serum sample using Quantikine ® ELISA Human Cystatin C Immunoassay (R&D Systems, Minneapolis, MN). Optical density was measured by a microplate reader (Synergy HT, BioTek Instruments, Inc., Winooski, VT) set to 450 nm. The limit of detection was 31 pg/ml. Creatinine was measured in serum/plasma samples (previously stored at −70 C) using Creatinine FS reagent and Respons 910 (both by DiaSys, Holzheim, Germany), based on Jaffe’s kinetic method without deproteinization, with the detection limit of 0.2 mg/dL.

eGFR Calculations:

Estimated glomerular filtration rate (eGFR) was calculated using cystatin C and/or creatinine according to multiple previously validated formulas. These included the Schwartz equation²⁷, wherein ${\text{eGFR}}_{\text{Schwartz}}=\frac{k*\text{height}\text{in}\text{cm}}{S\text{Cr}\text{in}\text{mg}\text{/dL}}$ (k=0.7 for adolescent boys, and 0.55 for adolescent girls, and children (male or female) <13 years old), the 2012 cystatin C equation²⁸, wherein eGFR_CysC2012 = 70.69 ∗ (Cystatin C) ^−0.931 and cystatin C is in mg/L.

Covariates:

Child weight and standing height were measured using a professional digital scale (InBody230, InBody, Seoul, Korea) and stadiometer with the head in the Frankfort plane at the 8–12 year visit. Weight and height were used to calculate BMI categories according to the WHO z-scores for age and sex, defined as obesity (z-score greater than 2), overweight (z-score between 1 and 2), and normal weight (z-score between −2 and 1). Participants in the underweight category (z-score below −2) were excluded (n=2). Information on maternal age, socioeconomic status, and environmental tobacco exposure were collected at the 2^nd trimester visit using a standardized questionnaire. Household environmental tobacco smoke exposure was dichotomized as yes/no based on the mother’s report that at least one household member smoked during the pregnancy. The SES index during pregnancy was calculated based on the 1994 Mexican Association of Intelligence Agencies Market and Opinion (AMAI) rule 13 * 6. The index classifies families into 6 levels based on 13 questions related to the characteristics of the household. The majority of families were middle or low SES. Thus, the 6 resultant levels were further collapsed into 3 SES categories: lower, medium, and higher. Gestational age at delivery was calculated using maternal report of last menstrual period (LMP) and confirmed with Capurro physical examination at birth.²⁹ When the physical examination assessment of gestational age differed by more than 3 weeks from the gestational age based on LMP, the physical exam was used in lieu of the LMP-determined gestational age.

Statistical Analyses:

BLLs were transformed via natural logarithm to approximate a normal distribution. Bone lead concentration was used on its original scale. Covariates in adjusted models (age, sex, socioeconomic status, BMI z-score, secondhand tobacco exposure) were selected a priori. Potential influential data points were assessed using Cook’s distance. We used multiple linear regression models to assess the association between each of the lead exposure biomarkers and the outcome variables. The outcome variables included BMI z-score (to assess direct effects with obesity) as well as eGFR_Schwartz and eGFR_CysC2012. For simplicity, we present results with eGFR_CysC2012 in the main tables. Preadolescent BMI categorization thresholds were defined according to WHO recommendations. We performed two sensitivity analyses that excluded participants 1) born preterm (less than 37 weeks) or 2) low birth weight (less than 2.5 kg). When there was evidence that the variance in eGFR increased with the mean, quasi-likelihood estimation was used resulting in a scaled Poisson objective function, and results agreed with linear regression models (data not shown).³⁰ All analyses were performed using R Studio version 3.6.2.

Results:

Table 1 shows the demographic characteristics of the 453 mother-child pairs with complete data. Participants were 8 to 12 years of age, and 50% were girls. A majority of participants (53%) were lower SES, and 29% of mothers reported exposure to indoor secondhand tobacco smoke in the home. Approximately half were normal weight, and 24% and 23% were considered overweight or obese respectively. The average serum creatinine was 0.43 mg/dL, which is within the expected range for this age.³¹ The eGFR_CysC2012 values were within the normal physiological range, with a median value of 97.1 mL/min/1.73m², and four were below 60 mL/min, a level associated with adult CKD.³² Maternal 2^nd trimester BLLs ranged between 0.7 and 17.8 μg/dL, and nearly 20% of maternal 2^nd trimester BLLs were above the CDC guidance level of 5 μg/dL, consistent with prior studies in the PROGRESS cohort.^11,12

Table 1:

Participant demographics for n=453 child-mother pairs in the PROGRESS cohort.

	Arithmetic mean (SD) or median [IQR]	N
Age, years	9.6 [9.2, 10.2]	453
Male, %	50.1%	453
Socioeconomic status		453
Lower	52.5%
Medium	37.5%
Higher	9.9%
Exposure to secondhand smoke (yes/no)	28.7%	453
Preadolescent BMI		453
Underweight	0.4%
Normal Weight	52.8%
Overweight	23.8%
Obese	23.0%
Cystatin C (mg/L)	0.71 [0.6, 0.8]	453
BUN (mg/dL)	12.0 [10.0, 13.7]	367
Creatinine (Plasma) (mg/dL)	0.44 (0.08)	448
Creatinine (Serum) (mg/dL)	0.43 (0.1)	368
eGFRSchwartz (Plasma) (mL/min/1.73m2)	130.8 [115.2, 148.7]	448
eGFRSchwartz (Serum) (mL/min/1.73m2)	173.4 [150.4, 204.4]	368
eGFRCysC2012 (mL/min/1.73m2)	97.1 [83.4, 111.8]	453
Prenatal Pb levels:
Second trimester blood Pb (μg/dL)	2.9 [2.0, 4.5]	453
Third trimester blood Pb (μg/dL)	3.1 [2.0, 4.8]	375
Cord Blood Pb (μg/dL)	3.3 [2.1, 5.3]	362
Tibia Pb	3.3 [−1.7, 7.8]	356
Patella Pb	3.4 [−1.2, 9.0]	353

SD, standard deviation; IQR, interquartile range; BMI, body mass index; BUN, blood urea nitrogen; eGFR_CysC2012, estimated glomerular filtration rate from 2012 cystatin C equation.

We first examined relationships between eGFR and important covariates, including age and BMI. In models adjusted for sex, BMI z-score, indoor tobacco smoke exposure, and socioeconomic status, a single year increase in age was significantly associated with an 8.6-unit increase in eGFR_CysC2012 (95% CI: 5.45, 11.8; p < 0.0001). As BMI is a significant risk factor for CKD, we also assessed the relationship between continuous BMI z-score and eGFR, and also stratified according to WHO-established BMI categories. We observed a marginal relationship between continuous BMI z-score and eGFR_CysC2012 (β_unadjusted = −1.37, 95% CI: −3.0, 0.26, p = 0.1) (Supplemental Figure 1). We also observed a null relationship between raw BMI and eGFR_CysC2012 (β_unadjusted = −0.39, 95% CI: −0.95, 0.18, p = 0.2). Stratified by BMI category, we observed a null relationship among normal weight and overweight participants (Supplemental Figure 2). However, among participants with obesity, we observed a significant 5.9-unit decrement in eGFR_CysC2012 (β= −5.9, 95% CI: −11.0, −0.8, p = 0.02). We also observed null relationships between measures of bone and blood lead levels with preadolescent continuous BMI z-score (data not shown).

Null overall and sex-specific relationships between prenatal lead exposure and preadolescent eGFR

Overall, we observed null relationships between measures of bone and blood lead and preadolescent eGFR (Table 2). To evaluate whether there was evidence of sex-specific effects, we conducted analyses stratified by boys and girls (Table 2). Among girls (n=226), we observed null relationships between all measures of lead and eGFR. Among boys (n=227), we observed null relationships between 2^nd trimester, 3^rd trimester, cord blood at delivery, and tibia lead levels and eGFR. We observed a significant, inverse association between patella lead levels and eGFR among boys, wherein a one-unit increase in patella lead was associated with a 0.4-unit decrease in eGFR_CysC2012 (95% CI: −0.8, −0.03, p = 0.03).

Table 2: Associations between overall prenatal lead exposure and preadolescent eGFR_CysC2012 at age 8 to 12 years, as well as sex-specific associations.

Estimates were adjusted for BMI z-score, SES, tobacco exposure, age, and sex (sex excluded from sex-stratified models).

	All Participants			Male Participants			Female Participants
	Estimate (95% CI)	P-Value	n	Estimate (95% CI)	P-Value	n	Estimate (95% CI)	P-Value	n
2T BLL	−0.81 (−4.1, 2.5)	0.63	453	−2.25 (−7.03, 2.53)	0.35	227	0.79 (−3.82, 5.41)	0.73	226
3T BLL	0.27 (−3.2, 3.8)	0.88	375	2.78 (−2.49, 8.06)	0.30	182	−0.71 (−5.53, 4.11)	0.77	193
Cord BLL	−2.14 (−5.8, 1.5)	0.25	362	−3.27 (−8.58, 2.05)	0.23	188	−1.10 (−6.38, 4.18)	0.68	174
Tibia Pb	0.07 (−0.2, 0.3)	0.63	356	0.12 (−0.25, 0.50)	0.52	171	0.02 (−0.39, 0.44)	0.91	185
Patella Pb	−0.15 (−0.4, 0.1)	0.27	353	−0.42 (−0.81, −0.03)	0.03	174	0.06 (−0.31, 0.44)	0.74	179

2T, second trimester; 3T, third trimester; BLL, blood lead level after ln-transformation.

Effect modification by BMI

When stratified by BMI status, we observed null relationships between all measures of lead and eGFR in the normal weight and obese groups (Table 3; Figure 1). However, among overweight participants, we observed significant, inverse associations between 2^nd trimester and 3^rd trimester BLLs as well as cord blood at delivery and children’s eGFR (Figure 1). For example, among overweight participants, a one ln-unit increase in 2^nd trimester BLL was associated with a 10.5-unit decrease in eGFR_CysC2012 (95% CI: −18.1, −2.8; p = 0.008). Similarly, among overweight participants, a one ln-unit increase in 3^rd trimester BLL was associated with a 10.0-unit decrease in eGFR_CysC2012 (95% CI: −17.6, −2.4; p = 0.01). Among overweight participants, a one ln-unit increase in cord BLL was associated with an 8.9-unit decrease in eGFR_CysC2012 (95% CI: −16.8, −0.9, p = 0.03). We observed evidence of multiplicative interaction between overweight status and second trimester BLLs among all participants (p = 0.003). We also observed no significant interaction between obesity status and second trimester BLLs (p=0.9).

Table 3: Associations between prenatal lead exposures and preadolescent eGFR_CysC2012, stratified according to BMI status.

Estimates were adjusted for SES, tobacco exposure, age, and sex.

	Normal weight			Overweight			Obese
	Estimate (95% CI)	P-value	n	Estimate (95% CI)	P-value	n	Estimate (95% CI)	P-value	n
2T BLL	1.76 (−2.4, 6.0)	0.42	239	−10.49 (−18.1, −2.8)	0.008	108	1.66 (−5.9, 9.2)	0.67	104
3T BLL	2.94 (−1.5, 7.4)	0.20	196	−9.97 (−17.6, −2.4)	0.01	95	4.29 (−4.7, 13.2)	0.34	82
Cord BLL	1.94 (−2.9, 6.8)	0.43	195	−8.87 (−16.8, −0.9)	0.03	82	−6.96 (−15.9, 2.0)	0.13	84
Tibia Pb	0.19 (−0.2, 0.6)	0.29	195	−0.13 (−0.8, 0.5)	0.69	84	−0.06 (−0.7, 0.6)	0.84	75
Patella Pb	−0.12 (−0.4, 0.2)	0.47	194	−0.52 (−1.3, 0.3)	0.22	81	−0.16 (−0.8, 0.5)	0.60	76

2T, second trimester; 3T, third trimester; BLL, blood lead level after ln-transformation.

Figure 1:

Association between second trimester BLL and eGFR_CysC2012 stratified by normal weight (n=239), overweight (n=108), and obese (n=104) as defined by WHO age- and sex- standardized BMI z-score.

Sensitivity Analyses

In a sensitivity analysis that excluded participants born preterm (n=45), we observed results similar to our main findings (Table 4). We found a significant inverse association between 2^nd trimester and 3^rd trimester BLLs and eGFR among overweight participants (Supplemental Figure 3). A one ln-unit increase in 2^nd trimester BLL was associated with an 8.4-unit decrease in eGFR_CysC2012 (95% CI: −16.7, −0.1, p = 0.05). A one ln-unit increase in 3rd trimester BLL was associated with an 8.9-unit decrease in eGFR_CysC2012 (95% CI: −16.9, −0.8, p = 0.03). We observed a marginally significant inverse association between cord BLL and eGFR, such that a one ln-unit increase in cord BLL at delivery was associated with a 7.4-unit decrease in eGFR_CysC2012 (95% CI: −15.9, 1.1, p = 0.09). We observed null relationships between all measures of lead (blood and bone) and eGFR in the normal weight and obese strata. Although not significant, one ln-unit increase in cord BLL was associated with an 8.3-unit decrease in eGFR_CysC2012 (95% CI: −16.9, 0.35, p=0.06) among participants with obesity.

Table 4: Sensitivity analyses excluding children born preterm (n=45) or low birth weight (n=36): Associations between prenatal BLLs (μg/dL) and preadolescent eGFR_CysC2012.

Estimates were adjusted for SES, tobacco exposure, age, and sex.

		Normal weight			Overweight			Obese
		Estimate (95% CI)	P-value	n	Estimate (95% CI)	P-value	n	Estimate (95% CI)	P-value	n
Term Participants	2T BLL	2.67 (−1.8, 7.2)	0.24	220	−8.43 (−16.75, −0.12)	0.05	96	−0.66 (−8.6, 7.3)	0.87	91
	3T BLL	2.84 (−1.8, 7.5)	0.23	183	−8.86 (−16.90, −0.82)	0.03	86	1.44 (−7.3, 10.1)	0.74	71
	Cord BLL	2.08 (−2.9, 7.1)	0.41	181	−7.39 (−15.88, 1.11)	0.09	73	−8.26 (−16.9, 0.35)	0.06	74
Normal Birth Weight Participants	2T BLL	1.58 (−2.7, 5.9)	0.47	220	−8.57 (−16.4, −0.7)	0.03	100	0.40 (−7.7, 8.5)	0.92	95
	3T BLL	2.50 (−2.0, 7.0)	0.28	182	−9.26 (−17.2, −1.3)	0.02	88	2.03 (−7.6, 11.6)	0.67	74
	Cord BLL	0.48 (−4.4, 5.3)	0.84	179	−8.22 (−16.2, −0.2)	0.04	77	−9.04 (−18.4, 0.3)	0.06	76

We performed an additional sensitivity analysis that excluded participants with low birth weight status (n=36). Some participants were both born preterm and at low birth weight and were therefore excluded from both sensitivity analyses. Excluding low birth weight participants, we similarly observed a significant, inverse association between 2^nd trimester, 3^rd trimester and cord BLLs and eGFR among overweight participants (Table 4; Supplemental Figure 4). A one ln-unit increase in 2^nd trimester BLL was associated with an 8.6-unit decrease in eGFR_CysC2012 (95% CI: −16.4, −0.7, p = 0.03). A one ln-unit increase in 3^rd trimester BLL was associated with a 9.3-unit decrease in eGFR_CysC2012 (95% CI: −17.2, −1.3, p = 0.02). A one ln-unit increase in cord BLL was associated with an 8.2-unit decrease in eGFR_CysC2012 (95% CI: −16.2, −0.2, p = 0.04). Again, we observed null relationships between all measures of lead (blood and bone) and eGFR in the normal weight and obese strata. Although not significant, one ln-unit increase in cord BLL was associated with a 9.0-unit decrease in eGFR_CysC2012 (95% CI: −18.4, 0.3, p=0.06) among participants with obesity.

Discussion:

In this study, we found that among overweight children, an increase in BLL at multiple timepoints in pregnancy and at delivery was associated with a significant decrease in eGFR, suggesting that low-level prenatal lead exposure may have important implications for kidney health in adolescence in the context of childhood adiposity. Given the near global epidemic of child overweight and obesity and the ubiquity of low-level lead exposure worldwide, if confirmed in other studies, the findings may have important public health implications for mitigating the burden of CKD and maintaining a normal weight through childhood.

There are several important observations from this work. One appropriate interpretation of this data is that preadolescent BMI modifies the effect of prenatal lead exposure on kidney function by the end of the first decade of life. Because the results depend on assessing interaction, it may explain why previous studies have demonstrated conflicting results regarding the effect of early life lead exposure on children’s kidney function: both factors must be present. For example, in a prospective study of 117 Bangladeshi children 4–5 years of age, an increase of 85 μg/kg in erythrocyte lead levels at 30 weeks’ gestation was associated with a 6.0 cm³/m² decrease in kidney volume after adjusting for potential confounders.³³ The authors found no evidence of effect modification by BMI z-score, although comparison is limited by the difference in lead biomarker, younger study population age,³³ and the much lower prevalence of child obesity in Bangladesh.³⁴ Indeed, child obesity is more prevalent in middle income and developed countries. Also, in a cross-sectional analysis of a prospective cohort of 391 Canadian and American participants aged 1–19 years of age, BLLs were associated with a decrease in eGFR, although the finding did not reach statistical significance.³⁵ Child obesity is less common in children under five³⁶ and lead exposure is less common in older children, which may have obscured evidence of interaction in prior studies with a wide age range. In the same North American study, subgroup analyses that stratified by anemia status or glomerular disease status found significant, inverse associations between BLLs and eGFR. Other cross-sectional studies have assessed the effects of lead exposure with other kidney parameters including proteinuria and observed mixed results.³⁷

Potential mechanisms of lead-induced nephrotoxicity include the production of reactive oxygen species and a weakened antioxidant response.^38,39 Although not well understood, manifestations of acute lead nephrotoxicity include glomerular and proximal tubule dysfunction, alterations in mitochondrial structure and energy production, and formation of nuclear and cytosolic inclusion bodies.⁴⁰ Prenatal lead exposure has been shown to alter genomic DNA methylation and microRNA expression^15,41 that could impact fetal nephrogenesis.⁴² Additionally, changes in DNA methylation patterns in the kidney cortex have been shown to influence glomerular filtration rate (GFR) and kidney fibrosis.⁴³

The interplay between lead exposure and early life metabolic health may be in part explained by lower nephron endowment coupled with alterations in the renin-angiotensin-aldosterone system. Fetal nephrogenesis is subject to modification according to the energy made available to the fetus.⁴⁴ Prenatal lead exposure can disrupt mitochondrial energy production⁴⁵ potentially resulting in decreased nephron number at birth, which carries an increased risk for CKD later in life.⁴⁴ Furthermore, lead can increase angiotensin-converting enzyme activity in rats as well as renin and aldosterone concentrations in both humans and animals^38,46, which could compound obesity-associated kidney decline. By driving a greater increase in renin and aldosterone concentrations, lead may worsen hypertension and glomerular hyperfiltration, leading to more rapid kidney decline.⁴⁷

However, effect modification by BMI does not completely explain the counterintuitive null results observed when stratified by children with obesity. There may be additional factors that are not accounted for, and there is a complex relationship between metabolic and renal health. For example, among adults, increased adiposity acts as a risk factor for de novo CKD while also conferring survival advantages in advanced CKD, a phenomenon known as the obesity paradox.⁴⁸ Despite consistent data, the mechanisms of this paradox are unknown and it is unclear whether the obesity paradox may present in adolescents with obesity. It is also worth noting that in sensitivity analyses that excluded both preterm or low birth weight participants, we observed marginally significant inverse results between cord blood lead levels and eGFR among the obese stratum. Furthermore, it is possible that eGFR estimating equations that use either cystatin C or creatinine are less appropriate for children with higher BMI. This could explain the lack of association between lead and eGFR in the obese group. The greater level of uncertainty in eGFR among individuals with higher BMI is evidenced by studies in adults^20,49, and in this study, we observed larger confidence intervals among overweight and obese participants (Supplemental Fig 2). Finally, it is possible to conclude that the lack of association between lead and eGFR in the obese stratum conflicts with the conclusion that BMI is an effect modifier, and there may be additional unmeasured confounding contributing to the observed associations. Overall, our findings may underscore the importance of maintaining a healthy weight through adolescence to promote kidney health.

Our study has many strengths. PROGRESS is an established prospective birth cohort with well-characterized demographic and covariate data, as well as simultaneous exposures. Maternal lead was measured years prior to eGFR, and therefore cannot be biased with respect to eGFR. We assessed lead in two biological matrices highly relevant to the prenatal period: 1) bone, more reflective of cumulative lead exposure and 2) blood level, more reflective of more recent, active lead exposure.²⁶ The lead biomarkers have relatively strong correlation, which has been described previously.⁵⁰ Our findings of consistent null relationships between bone lead and eGFR, even when stratified by BMI, suggest that blood lead is more biologically active and therefore has a more significant role in impaired kidney function. We assessed eGFR using three common equations, including the recently updated cystatin C equation for children.²⁸ Decrements in kidney function have been historically assessed by biomarkers such as serum creatinine and urea.⁵¹ However, changes in nearly all of these biomarkers may not manifest until substantial kidney impairment is present or clinical kidney disease is apparent. For example, after accounting for risk based on serum creatinine concentration in children with CKD, there is a wide variation in CKD progression.⁵² We note that models using the Schwartz formula eGFR did not replicate the results of models that used cystatin C-based eGFR. The cystatin C univariate equation for children has advantages over these estimates, and eGFR calculations that are based on the metabolite cystatin C have been proposed as an improved assessment that more closely estimates the gold standard of measured GFR in pediatric populations.^49,53 Cystatin C is produced by all mononucleated cells and freely filtered at the glomerulus. Further, compared to serum creatinine, cystatin C is not as subject to variation with muscle mass, BMI and delayed sensitivity to kidney damage.⁵⁴ While differences between the two eGFR equations could be attributed to the differences in biomarker, the Schwartz formula would be expected to trend in the same direction. Another possible explanation for the variations is that serum creatinine was measured using the Jaffe method standardized according to the reagent provider, and was not calibrated via isotope dilution mass spectrometry.⁵⁵

Our study has limitations. We assessed eGFR at a single time point between 8 and 12 years of age, which does not capture the adolescent growth spurt, although longitudinal follow-up is ongoing and will include adolescence. The PROGRESS cohort is also ethnically homogeneous, which limits the extent to which this analysis is externally generalizable but strengthens its internal generalizability by reducing confounding due to genetics or social factors. In addition, the WHO BMI Z-score reference population is American, and may not be generalizable to our cohort of Mexican children.⁵⁶ As this is an observational study, we are also unable to report a specific mechanistic pathway between prenatal lead exposure, increased adiposity, and decreased preadolescent glomerular function. As increased adiposity can cause glomerular hyperperfusion and hypertrophy followed by kidney function decline, it is also possible that there is greater heterogeneity of eGFR at a single timepoint among participants with obesity. The hyper- vs. hypofiltration state may be dependent on the length of time an individual has been affected by obesity rather than BMI at a single time point. Longitudinal assessments of the relationship between obesity and kidney function trajectories through adolescences are needed.

Conclusion

In conclusion, our findings suggest a complex relationship between preadolescent BMI and prenatal lead exposure, in which the two may act together to worsen kidney function later in life. These findings underscore the public health need to prevent lead exposure in pregnancy as well as childhood obesity. To better understand the complex relationship between environmental toxic metal exposure and the origins of CKD, future longitudinal assessments of nephrotoxicant exposure and preadolescent kidney function will improve our understanding of risk factors for kidney impairment and associated comorbidities.

Highlights:

• Overall, we observed null associations between perinatal lead exposure and kidney function

• Perinatal BLLs were associated with reduced preadolescent eGFR among overweight participants

• We describe possible effect modification by adiposity status

• The complex relationship between lead exposure and adiposity warrants further study