Early childhood lead exposure and adolescent heart rate variability: A longitudinal cohort study

Abstract

Lead is a known neurotoxicant with many detrimental health effects, including neurocognitive deficits and cardiovascular and metabolic disorders. However, few studies have tested the association between lead exposure and the physiological stress response, which in and of itself may act as a precursor to and/or underlying mechanism of detrimental health outcomes. The purpose of this study was to examine the influence of early childhood and early adolescent low-level lead exposure on early adolescent heart rate variability, a widely-used measure of physiological stress. Participants were 408 children from Jintan, China for whom blood lead levels were measured between 3-5 years (early childhood) and again at 12 years (early adolescence). Heart rate variability was assessed at 12 years while participants underwent an induced stress task utilizing the ratio of low to high frequency (LF/HF) ECG measures. Mean blood lead levels in the cohort were 6.63 mcg/dl and 3.10 mcg/dl at 3-5 years and 12 years, respectively. Blood lead levels at 3-5 years of age (β 0.06, p= 0.027), but not at age 12 (β −0.05, p=0.465), were significantly associated with LF/HF measures while controlling for multiple sociodemographic variables, potentially reflecting a dysregulated stress response with a shift towards sympathetic dominance. These findings suggest that early childhood lead exposure may have a detrimental influence on early adolescent autonomic responses to acute stress, which holds implications for cardiovascular health and overall growth and development.

1. Introduction

While many large-scale sources of lead exposure have been reduced, adults and children are still routinely exposed to lead via air pollution, electronic waste, lead pipes, ceramic glazes, and even baby food (1-3). Exposure to lead during childhood and adolescence is associated with a host of detrimental outcomes that persist into adulthood (4). In fact, lead exposure during these critical developmental periods has been associated with long-term impairments to neurobehavioral function as well as increased risk for detrimental cardiometabolic outcomes in many studies (5-10) though not all (11-13). It is, therefore, possible that lead exposure in the childhood developmental years may also impact long-term biological functioning.

Physiological stress response is a measure of biological functioning which has been associated with lead exposure in few studies of adults (14, 15) and children (16, 17). Both the hypothalamic-pituitary-adrenal (HPA) axis and autonomic nervous system (ANS) are responsible for mounting an appropriate response to external stressors. However, interference with these system’s negative feedback loop induces a dysregulated physiological stress response and impairs the return to basal functioning levels (18). The result is physiological dyshomeostasis, where body systems are constantly “ramped up” and unable to respond appropriately to stressors.

Heart rate variability (HRV) in response to external stressors is an important biomarker of physiological stress responses, thought to represent the balanced functionality of the sympathetic and parasympathetic nervous systems, and has been widely used to assess ANS function because it is non-invasive and reliable. (19). While there is no universally accepted measure to assess physiological stress, HRV has been widely recognized as a valid measure of the construct (20). HRV should increase in response to a stressor. Conversely, decreased HRV is indicative of a dysregulated stress response, unbalanced sympathetic and parasympathetic output, and suggests long-term and potentially permanent changes to the ANS resulting in dyshomeostasis (20). Dysregulated HRV has been associated with other biomarkers of dyshomeostasis, such as cortisol at awakening levels and during induced stress tests, as well as self-reports of chronic stress in children (20-23).

Few studies with children examine the association of childhood lead exposure and physiological stress responses. Gump and colleagues (2008) report a positive association between prenatal and postnatal blood lead (mean <10mcg/dl) and reactive salivary cortisol in 9.5 year old children (24). Considering autonomic function, blood lead (mean 1.01 mcg/dl) has been cross-sectionally associated with inhibition of the sympathetic and parasympathetic response to acute stress in 9-11 year old children (16). However, others report null cross-sectional associations between very low blood lead levels (0.98 mcg/dl) and HRV indices in 11 year old children (17).

Further research is needed in child and adolescent populations with low-level lead exposure and HRV assessment occurring during these periods to establish the relationship between lead exposure and stress response physiology. Childhood and early adolescence are considered critical periods for development, during which the brain is highly plastic and neural structures instrumental in physiological stress responses are maturing (25). During this time the brain is particularly susceptible to environmental insults (26). Moreover, these are time periods when the sequelae of stress response system dysfunction are more likely to impair longer-term neurobiological and broader health outcomes (27). Utilizing a large community sample of Chinese children, this study examines both the long-term and cross-sectional relationships between lead exposure at two critical child developmental periods, at 3-5 and 12 years of age, and an HRV frequency measure in response to an acute stressor at 12 years of age. We hypothesized that lead exposure will be positively related to the HRV frequency measure, indicating a decreased overall HRV.

2. Materials and Methods

2.1. Design

This study is part of an ongoing, longitudinal observational project, the China Jintan Child Cohort Study. This project includes a preschool cohort with three levels: bottom (3-4 years old), middle (4-5 years old), and top (5-6 years old). Between the Fall of 2004 and Spring 2005, children and their parents were recruited from four preschools in Jintan city, Jiangsu province, China. Preschools were chosen to represent the city’s geographic, social, and economic profiles (28). In this population, a major source of lead exposure comes from dust and soil, which are contaminated through pollution from traffic and industrial waste sites (29). 1656 children were initially recruited, with 97% agreeing to participate. 1385 children participated in the first wave of data collection (Wave 1, Figure 1). General information from children and parents was collected during school hours, whereas physiological data was collected in a laboratory setting. Between 2011-2013 children were invited back to participate in a second wave of data collection (Wave 2), when they were in 6^th grade and approximately 12 years of age (N=1100). More detailed information on recruitment and enrollment procedures is reported in a cohort profile update (30, 31).

Figure 1.

Jintan China Child Cohort recruitment and follow up flow chart

2.2. Participants

For these analyses, we included a subsample of children (Figure 1, N= 408) from the 1,100 who participated in the second wave of data collection and had complete data on blood lead concentrations, HRV outcomes, and all covariates (age at HRV testing, sex, parental occupation, neighborhood conditions, and serum Fe levels). While all children in Wave 2 were invited to participate in HRV testing, many declined (N= 534). A majority of declinations were due to parental time constraints. While HRV testing itself took one hour, there were other testing procedures scheduled for the same day included EEG and neurocognitive testing. Parents either did not have the time to accompany their children for many hours or, given academic pressure in this cultural setting, did not want their children missing school. Some also declined due to feeling uncomfortable with the ECG and EEG testing instruments, despite these being non-invasive measures. A select number of participants were removed after consenting to HRV testing for reasons including health problems, movement artifact, having very high or low heart rates, or technical issues (N=19), which resulted in 547 HRV observations. Written informed consent was obtained from parents and participants at both Waves 1 and 2. Institutional review board approval was obtained from the University of Pennsylvania and the Ethical Committee for Research at Jintan Hospital in China.

2.3. Measures

Blood lead levels at 3-5 years and 12 years.

Blood lead levels (BLLs) were collected twice for the participants. The first collection was when children were 3, 4, or 5 years of age, between November 2004 and March 2005, and again when children were approximately 12 years old between 2011-2013 in the months of November-March. Blood samples were obtained at school health clinics located at each elementary school attended by cohort children. Collection was done by a trained pediatric nurse utilizing a standardized research protocol to avoid lead contamination (32). Samples were frozen and shipped to the Research Center for Environmental Medicine of Children at Shanghai Jiaotong University. This laboratory participated in a CDC administered quality-control program (Blood Lead Proficiency Testing Program) for the measurement of lead in whole blood. Specimens were analyzed twice via graphite furnace atomic absorption spectrophotometer using a replication procedure with the final measure being the mean of repeated measurements (32). No observations were removed for high variability between measures. Blood lead reference materials for quality control (QC) were provided by Kaulson Laboratories, New Jersey, and QC samples were added blindly to the study samples. The limit of detection (LOD) was 1.8 mcg/dl for Wave 1 and 1.0 mcg/dl for Wave 2. Samples below the LOD were considered as the LOD (N=3). Further details on sample collection have been previously published (33).

HRV outcomes at 12 years.

Data on HRV was collected when participants were approximately 12 years old in the summer and fall between 2011-2013 in a controlled psychophysiology lab located on the diagnostic floor of the Jintan Hospital. Parents and children were given instructions to rest well the night prior to testing and avoid exercise, overeating, or stressful situations the day of testing. At arrival, children were asked to sit quietly for 3-5 minutes to allow for baseline relaxation. For additional relaxation time, general health questions were asked for approximately 15 minutes and children were given a popular cartoon movie to watch prior to testing.

Children then underwent a standardized stressful stimulus, the Public Speaking Stress task, which has been recognized as a self-relevant stressor (34). This task required participants to first plan a speech to deliver and then deliver the speech to the research assistant, yielding two measurement timepoints. Trained research assistants collected HRV measurements during each phase of the task using non-invasive electrocardiographs (ECG) that were recorded axially on the left and right ribs at the level of the heart (to avoid movement artifact) using silver/silver chloride adhesive disposable electrodes. In accordance with published protocols, data were recorded using a bandpass of 0.5-35 Hz and a 50 Hz notch filter, and the recording digitized at 1000 Hz (35). ECG data were cleaned for artifacts manually using AcqKnowledge analytic tools to identify unusually large changes in heart rate. HRV was derived by first extracting the interpolated tachogram (a continuous time-domain representation of R-R intervals, the time elapsed between R waves on a QRS signal from the ECG) from the ECG data in AcqKnowledge and then performing a fast-Fourier transform in MATLAB to average power into frequency bands of low frequency, 0.04-0.15 Hz (LF) and high frequency, 0.15-0.4 Hz (HF). Power in each frequency band was averaged for both phases of the stress task. The ratio between LF and HF power bands (LF/HF), which is thought to operationalize parasympathetic and sympathetic balance, was used to assess physiological stress responses (22, 36). An increasing LF/HF ratio corresponds to a shift to sympathetic dominance and decreasing HRV, suggesting a dysregulated stress response, with the normal range being 1.5 ± 2.0 (ms²) (37). Detailed methodology for heart rate measurement is given elsewhere (38).

Covariates.

Covariates were selected based on known predictors of BLLs in our data set (39). First, neighborhood conditions, defined as having a large number of neighbors, were included. This was a subjective measure answered by the parents when asked ‘does the place where you live have many neighbors?’. Parents would indicate whether or not this was true for them, and the measure was dichotomized as a yes or no. Second, self-reported parental occupation (unemployed, skilled/unskilled, and professional), a proxy for socioeconomic status, was included as a covariate. Both neighborhood conditions and parental occupation were assessed via a parental questionnaire administered in Wave 1. Additional covariates were selected due to potential associations with HRV outcomes (40, 41). These included sex and age at HRV collection which were gathered via a child questionnaire administered in Wave 2. Finally, serum Fe levels were included as a covariate due to postulated relationships between Fe deficiency and increased lead absorption (42). Additionally, Fe deficiency has been shown to influence HRV, further warranting it’s inclusion as a covariate (43). Fe samples were collected venously in a lead-free EDTA tube for Fe analysis during Wave 1. Samples were frozen at −20 °C and shipped to the Child Development Center, Nanjing Medical University, Nanjing, China, for analysis. Fe levels were determined by atomic absorption spectrophotometry (BH model 5.100 manufactured by Beijing Bohu Innovative Electronic Technology Corporation), with duplicate readings taken with an integration time of 2s.

2.4. Statistical analysis

Sample characteristics were summarized by descriptive statistics such as means, medians, standard deviation (SD), and frequencies. We inspected the data for normality and potential outliers. The differences between the included and excluded samples were compared using t tests and chi-squared tests. Pearson correlations were used to assess the bivariate relationships between BLLs and the HRV frequency measures. General linear models with robust standard errors were then used to examine associations between BLLs at each time point and the planning and speaking phase HRV frequency measures in models adjusted for parental occupation, neighborhood conditions, sex, age at HRV measurement, and serum Fe levels. For each phase HRV frequency measure, we computed two models, one with BLLs at 3-5 years as a predictor and one with BLLs at 12 years as a predictor. Sensitivity analyses were also run with both measures of BLLs in the model to confirm findings. Sex was a significant covariate in our models and, as such, we performed a further analysis with an interaction term between BLLs and sex entered into the model. As a further analysis, we dichotomized BLLs at 3-5 years at the clinical reference value established by the Centers for Disease Control and Prevention (CDC), ≤5 mcg/dl, >5 mcg/dl, and a series of general linear models with robust standard errors were conducted with the dichotomized BLLs predicting both the planning and speaking phase HRV frequency measure (44). As a sensitivity analysis, the BLL and HRV measures were log transformed and all analyses were rerun to check for consistency. All analyses were conducted with StataCORP 15 statistical package.

3. Results

3.1. Sample characteristics

The sample consisted of 219 males and 189 females with a mean age of 11.52 years at the time of HRV testing (Table 1). Within the HRV data, 3 high outlier observations (those with values greater than 6 standard deviations from the mean) were excluded. All variables were normally distributed, however, we repeated all analyses with log transformed BLL and HRV variables, which were consistent with those reported here. Mean BLLs at each time measure were 6.63 mcg/dl and 3.10 mcg/dl at 3-5 years and 12 years, respectively. Most mothers and fathers worked as unskilled or skilled laborers (Fathers: 59.07%, Mothers: 46.81%). The 408 children included in the analysis did not differ from those who were excluded in terms of age, sex, neighborhood conditions, BLLs, serum Fe levels, or father’s occupation. However, those excluded from the study did have a greater proportion of mother’s working in professional settings compared to those included in the study.

Table 1.

Participant characteristics

	Mean ± SD or number and (%) Median (IQR)		p-value 1
	Included (N= 408)	Excluded (N=692)
Blood lead 3-5 years	6.63 ± 2.64	6.33 ± 2.64	0.059
	6.4 (4.9 - 8)	6.1 (4.6 - 7.6)
Blood lead 12 years	3.10 ± 1.17	3.12 ± 1.17	0.832
	2.9 (2.3-3.6)	29.9 (2.3-3.7)
Father’s Occupation			0.369
Unemployed	19 (4.66%)	33 (3.87%)
Unskilled/skilled labor	241 (59.07%)	477 (55.92%)
Professional	148 (36.27%)	343 (40.21%)
Mother’s Occupation			0.005
Unemployed	120 (29.41%)	218 (25.29%)
Unskilled/skilled labor	191 (46.81%)	363 (42.11%)
Professional	97 (23.77%)	281 (32.60%)
Age at HRV test	11.52 ±0.36	11.54 ±0.40	0.305
	11.48 (11.25-11.75)	11.51 (11.27-11.80)
Sex			0.439
Male	219 (53.68%)	546 (55.94%)
Female	189 (46.32%)	430 (46.32%)
Serum Fe	8.15 ± 0.82	8.12 ± 0.83	0.574
	8.06 (7.59-8.62)	8.06 (7.51-8.66)
Neighborhood Conditions			0.425
Crowded	183 (44.85%)	336 (47.32%)
Not Crowded	225 (55.15%)	374 (52.68%)
Planning Phase HRV LF/HF	1.91 ± 1.54	1.94 ± 1.65	0.844
	1.55 (0.79-2.51)	1.42 (0.73-2.60)
Speaking Phase HRV LF/HF	2.30 ± 1.43	2.27 ± 1.61	0.813
	1.97 (1.26-2.91)	1.80 (1.29-2.96)

¹Comparison tests included t tests and chi-square tests for continuous and categorical variables, respectively.

Notes. Planning and Speaking phase LF/HF are measures of heart rate variability, a physiological stress response, assessed at the two time points during the Public Speaking Stress tasks. A high ratio represents lower heart rate variability, which is a dysregulated stress response indicative of dysfunction.

BLLs: Blood lead levels; HF: high frequency power band; IQR: Interquartile range; LF: low frequency power band; LF/HF: low frequency/high frequency ratio

3.2. Bivariate associations between BLLs and HRV

BLLs at 3-5 years were positively correlated with the speaking phase HRV measure (r 0.1103, p = 0.0258) only, such that as BLLs increased so did the LF/HF ratio, indicating a dysregulated stress response. This correlation, however, was modest. There were no significant correlations between early adolescent BLLs and HRV frequency measures during either the planning or speaking phase.

3.3. Associations between BLLs and HRV

In the adjusted models, neither BLLs at 3-5 years (Table 2, R² 0.0145, β= 0.03, p = 0.219, 95% CI [−0.02, 0.09]) nor at 12 years (Table 2; R² 0.0120, β= −0.04, p = 0.459, 95% CI [−0.16, 0.07]) were significantly associated with the planning phase HRV LF/HF measure. For the subjectively more stressful speaking phase, BLLs at 3-5 years (Table 3, R² 0.0539, β= 0.06, p = 0.021, 95% CI [0.01, 0.12]), but not BLLs at 12 years (Table 3, R² 0.0426, β= −0.05, p = 0.431, 95% CI [−0.18, 0.08]) were significantly associated with the LF/HF ratio in the adjusted models, indicating that higher BLLs in early childhood predicted a dysregulated stress response in early adolescence.

Table 2.

BLLs influence on the planning phase HRV frequency measure (LF/HF)

Predictors
	Crude Model 11	Model 1 1	Crude Model 2 2	Model 2 2
	R2 0.0041	R2 0.0145	R2 0.0011	R2 0.0120
BLLs 3-5 years	0.04 [−0.02, 0.09] 3	0.03 [−0.02, 0.09]
BLLs 12 years		-	−0.04 [−0.16, 0.07]	−0.04 [−0.16, 0.07]
Father Occupation
Unemployed		−0.25 [−1.11, 0.62]		−0.26 [−1.12, 0.60]
Unskilled/skilled		−0.04 [−0.41, 0.33]		−0.05 [−0.42, 0.31]
Professional		Ref		Ref
Mother Occupation
Unemployed		0.08 [−0.39, 0.55]		0.09 [−0.38, 0.57]
Unskilled/skilled		−0.03 [−0.46, 0.40]		−0.02 [−0.45, 0.41]
Professional		Ref		Ref
Age at HRV test		−0.14 [−0.51, 0.23]		−0.13 [−0.49, 0.23]
Sex
Female		0.09 [−0.21, 0.38]		0.07 [−0.23, 0.36]
Male		Ref		Ref
Serum Fe		0.15 [−0.02, 0.32]		0.16 [−0.02, 0.33]
Neighborhood
Crowded		−0.03 [−0.32, 0.26]		−0.02 [−0.30, 0.26]
Not Crowded		Ref		Ref

¹Model with BLLs at 3-5 years

²Model with BLLs at 12 years

³Coefficient [95% Confidence Interval]

Bolded= p-value < 0.05

Notes. Models represent the LF/HF outcome. An increasing LF/HF ratio indicates a decreasing HRV value.

Table 3.

BLLs influence on the speaking phase HRV frequency measure (LF/HF)

Predictors
	Crude Model 11	Model 1 1 R2	Crude Model 2	Model 2 2
	R2 0.0122	0.0539	R2 0.0023	R2 0.0426
BLLs 3-5 years	0.06 [0.004, 0.12] 3	0.06 [0.01, 0.12]		-
BLLs 12 years		-	−0.06 [−0.19, 0.07]	−0.05 [−0.18, 0.08]
Father Occupation
Unemployed		0.25 [−0.48, 0.98]		0.22 [−0.51, 0.96]
Unskilled/skilled		−0.01 [−0.34, 0.33]		−0.03 [−0.36, 0.30]
Professional		Ref		Ref
Mother Occupation
Unemployed		0.23 [−0.17, 0.63]		0.25 [−0.15, 0.65]
Unskilled/skilled		−0.01 [−0.39, 0.37]		0.01 [−0.36, 0.39]
Professional		Ref		Ref
Age at HRV test		0.17 [−0.23, 0.56]		0.18 [−0.22, 0.58]
Sex
Female		0.49 [0.21, 0.77]		0.46 [0.18, 0.74]
Male		Ref		Ref
Serum Fe		0.11 [−0.05, 0.26]		0.13 [−0.03, 0.28]
Neighborhood
Crowded		0.15 [−0.14, 0.43]		0.17 [−0.11, 0.45]
Not Crowded		Ref		Ref

¹Model with BLLs at 3-5 years

²Model with BLLs at 12 years

³Coefficient [95% Confidence Interval]

Bolded= p-value < 0.05

Notes. Models represent the LF/HF outcome. An increasing LF/HF ratio indicates a decreasing HRV value.

To further understand the relationship between 3-5 year BLLs and the HRV outcomes, we dichotomized the 3-5 year BLLs at the clinical reference value of 5 mcg/dl and conducted a series of linear regressions. Children with BLLs >5 mcg/dl had significantly greater speaking phase LF/HF ratio values compared to those with BLLs ≤5 mcg/dl, indicating sympathetic dominance shift and decreased HRV in those with greater early childhood lead exposure (Table 4, R² 0.0533, β= 0.36, p = 0.013, 95% CI [0.08, 0.65]). For the planning phase HRV frequency measure, however, there were no significant differences between the high and low BLL groups (Table 4, R² 0.0117, β= 0.10, p = 0.547, 95% CI [−0.22, 0.41]). As a further analysis, we examined potential effect modification of sex by including an interaction between BLLs and sex in our models. At 12 years, sex significantly moderated the relationship between BLLs and the speaking phase HRV outcomes (Table 5, R² 0.0595, β −0.32, p = 0.014, 95% CI [−0.58, −0.07]), such that as BLLs rose, males saw an increasing LF/HF ratio, and decreasing HRV, compared to females.

Table 4.

Mean planning and speaking HRV frequency measure (LF/HF) for children dependent on BLLs at age 3-5 years

Predictors
	HRV 11	HRV 1 1	HRV 2 2	HRV 2 2
	R2 0.0010	R2 0.0117	R2 0.0113	R2 0.0533
BLLs Dichotomized				-
> 5 mcg/dl	0.01 [−0.19, 0.41] 3	0.10 [−0.22, 0.41] 3	0.35 [0.06, 0.63]	0.36 [0.08, 0.65]
≤ 5 mcg/dl	Ref.	Ref.	Ref.	Ref.
Father Occupation
Unemployed		−0.25 [−1.12, 0.62]		0.27 [−0.47, 1.01]
Unskilled/skilled		−0.05 [−0.41, 0.32]		−0.01 [−0.34, 0.32]
Professional		Ref		Ref
Mother Occupation
Unemployed		0.09 [−0.38, 0.57]		0.23 [−0.16, 0.63]
Unskilled/skilled		−0.02 [−0.45, 0.41]		0.01 [−0.37, 0.38]
Professional		Ref		Ref
Age at HRV test		−0.12 [−0.48, 0.24]		0.20 [−0.19, 0.60]
Sex
Female		0.07 [−0.22, 0.37]		0.48 [0.20, 0.76]
Male		Ref		Ref
Serum Fe		0.16 [−0.02, 0.33]		0.15 [−0.03, 0.27]
Neighborhood
Crowded		−0.06 [−0.36, 0.24]		0.15 [−014, 0.43]
Not Crowded		Ref		Ref

¹Model with planning phase HRV LF/HF ratio

²Model with speaking phase HRV LF/HF ratio

³Coefficient [95% Confidence Interval]

Bolded= p-value < 0.05

Notes. Models represent the LF/HF outcome. An increasing LF/HF ratio indicates a decreasing HRV value.

Table 5.

BLL by sex interactions for HRV outcomes

BLLs x Sex 1	β [95% CI]
	Planning Phase HRV LF/HF (ms2)	Speaking Phase HRV LF/HF (ms2)
BLLs 3-5 years x Sex	0.01 [−0.10, 0.12]	0.02 [−0.09, 0.14]
BLLs 3-5 years	0.03 [−0.04, 0.10]	0.05 [−0.01, 0.12]
Female	0.01 [−0.72, 0.74]	0.34 [−0.40, 1.07]
BLLs 12 years x Sex	−0.08 [−0.32, 0.15]	−0.32 [−0.58, −0.07]
BLLs 12 years	0.00 [−0.18, 0.18]	0.13 [−0.07, 0.33]
Female	0.33 [−0.49, 1.15]	1.46 [0.61, 2.30]

¹Models adjusted for sex, age at HRV measurement, neighborhood conditions, parental occupation, and serum Fe, Bolded= p-value <0.05

4. Discussion

In this large childhood sample, the novel association between BLLs and early adolescent HRV, a physiological stress response and measure of ANS function, was found. In particular, we report marginal associations with BLLs during the planning phase of the stress task and significant associations during the more stressful speaking phase of the task, such that as BLLs increased, HRV appeared more dysregulated with an increasing LF/HF ratio. Further, children with BLLs >5 mcg/dl at 3-5 years had significantly greater LF/HF ratios compared to children with BLLs ≤5 mcg/dl, indicating greater stress response dysregulation and ANS dysfunction with greater BLLs. While these associations were significant, their effect size was small. However, as HRV is reflective of ANS functioning and, therefore, closely integrated with the cardiovascular, respiratory, neuroendocrine, immunological, and metabolic systems, these results suggest clinically relevant findings (21). To our knowledge, this is the first study to report a longitudinal effect of early childhood low-level lead exposure on HRV in early adolescence.

These results align with and expand upon a small body of literature suggesting an association between occupational lead exposure and dysregulated HRV. In previous studies of HRV outcomes in adults, there is a reported cross-sectional association between high levels of occupational blood lead (mean 20 mcg/dl) and decreased R-R intervals during deep breathing exercises, an outcome indicative of ANS dysfunction (45). Chronic occupational blood lead in male smelter workers (~ 25 mg/dl) was significantly associated with reduced HRV compared to controls (14). In older individuals with metabolic syndrome, bone lead (~19-23 mg/g) was similarly associated with reduced HRV (15). Our study suggests that even lower levels of lead exposure, in a child population with exposure during developmental time periods, are associated with similar ANS dysregulation as measured by HRV.

Our results confirm and expand upon the results from the few child studies examining autonomic functioning. Studies have previously focused on cross-sectional relationships in early adolescence. Gump and colleagues (16) report that in a sample of 140 children between ages 9 and 11, there is a cross-sectional association between BLLs (mean 1.01 mcg/dl) and ANS dysregulation evidenced through reduced high frequency HRV in response to an acute stressor, suggesting parasympathetic inhibition. However, a more recent examination of 203 children from the same cohort suggested no significant cross-sectional associations between BLLs (mean 0.98 mcg/dl) and LF/HF ratio HRV measures in 9-11 year old children (17). Our results similarly point to a lack of cross-sectional association between BLLs and HRV outcomes. However, we do report significant longitudinal associations between BLLs and the LF/HF frequency measure in a larger childhood sample, expanding on the current literature and aligning with outcomes reported in longitudinal investigations of other heavy metals (46).

Sex was only a significant predictor in our model with 12 year old BLLs predicting speaking phase HRV measures. Our interaction analysis suggested males saw increasing LF/HF ratios, and decreasing HRV, as BLLs rose compared to females. Previous research has suggested females overall have increased HRV compared to males (40). However, others report females having overall lower HRV measures compared to males, though females reported greater subjective stress than males (47). Our sample did not include measures of subjective stress which may be contributing to this significant interaction.

BLLs were only found to be significantly associated with the HRV frequency measure during the speaking phase of the Public Speaking Stress task, not during the planning phase. The speaking, versus planning phase, of the Public Speaking Stress task has been reported to elicit the greatest stress response in adults, where in significant changes in HRV were seen only during the speaking phase (48). However, other studies suggest the planning phase elicits the greatest changes in cardiovascular responses from baseline compared to when delivering the speech (34). Our study, however, is not able to account for baseline HRV in our participants and cannot measure changes in HRV from pre-task to intra-task.

These results suggest that lead exposure during early developmental years has a long-term impact on HRV, as early childhood BLLs (3-5 years) were more strongly predictive of early adolescent HRV than the cross-sectional relationship between BLLs and HRV at 12 years old. A potential reason for this finding is the mean BLLs at 3-5 years old were double that of the 12-year-old observations. While a gradual decline in BLLs as children age is an expected outcome frequently noted in population studies (49), it could be that higher levels of lead observed in our 3-5 year old sample are required to appreciate an effect on HRV outcomes. Another possible explanation for this finding is the timing and measurement of lead exposure in this sample. Our two collection points, in early childhood and early adolescence, are not representative of sustained lead exposure as the half-life of blood lead is only 1-2 months (50). Prenatally and across childhood, the brain is highly plastic and particularly sensitive to environmental insults such as lead exposure, and exposure at key timepoints may be more influential than others (4). Throughout these sensitive periods of development, some children in our sample may have experienced sustained, chronic lead exposure and others may only have experienced acute exposure, potentially influencing our results. While our study did not report associations between BLLs and HRV in adolescence, previous studies have found cross-sectional relationships between lead exposure and dysregulated stress response in early adolescence (16), which may be due to unreported prenatal and/or early childhood lead exposure. Studies on the influence of sustained lead exposure throughout developmental stages will help to determine the most susceptible periods of exposure.

The biological mechanisms underpinning the relationship between lead and physiological stress functionality are relatively unknown. Researchers suggest that the link between the paraventricular nucleus of the hypothalamus (PVN), where stress responses are activated, and lower-brain stem regions such as the nucleus of the solitary tract (NTS), where information from peripheral sensory receptors is integrated, could be serve as a mechanism (51). The PVN is a major sympoatoexcitatory region that coordinates the lower brainstem (i.e. NTS) to influence cardiovascular and respiratory activity in response to stress (52). Lead could act as a noxious agent within these individual areas, the PVN and NTS, or the pathway between them and modify the neuronal pattern of discharge by directly altering the nerve cells and/or indirectly by modifying the cell environment (51).

Dysregulated HRV is associated with a number of detrimental health outcomes, such as increased proinflammatory activity and cardiometabolic disease (53-55). Clinicians should be aware of this relationship and acknowledge a potentially greater risk of lead exposed individuals for detrimental health outcomes. It should be noted that the effect size of these relationship was relatively small, though clinically significant given links with detrimental health outcomes. Further, the mean BLL in this sample, which is still considered low, is higher than that of children in the United States where levels have dropped to ~0.89 mcg/dl in recent years (56). As the mean BLL in this sample was much higher than US children, we are unable to generalize these findings to the US or other populations of lower lead exposure.

4.1. Strengths and Limitations

Our study had several strengths, including the large community-based sample with robust, laboratory-controlled psychophysiological testing of HRV and evaluation of blood lead. In addition to biomarkers of the two key variables, blood lead and physiological stress, the study employed a longitudinal design with two waves of blood lead measurement across six years from early childhood to early adolescence and outcome measures assessed in Wave 2 allowing for temporal examination. Further, this sample includes overall low mean BLL measures, contributing to the low-level lead exposure literature.

However, there are several limitations that should be considered. First, our results from this observational study cannot confirm causality. However, temporal order was established with early lead measurements and later measures of physiological stress. Second, multiple unmeasured factors, such as diet, BMI, or physical activity could have confounded the relationship between BLLs and dysregulated stress responses. Season of BLL and HRV assessment could have also influenced these relationships (57, 58). Additionally, the parental occupation measure may not sufficiently capture the overall socioeconomic status of the child, which is known to influence stress functionality. Third, BLLs were only measured at two points in time and therefore represent acute as opposed to sustained lead exposure. Fourth, those included in the study differed from those excluded by mother’s occupational level by way of having a greater proportion of mother’s working professionally. The included sample may have been of overall lower socioeconomic status which may have influenced participants stress physiology and HRV outcomes. Finally, these data are from a region in China and, therefore, represent specific cultural outcomes. It is possible that these children differ significantly in their responses to the Public Speaking Stress task than children from other cultures. For example, in examining a Group Public Speaking task, researchers have found that children (ages 11-18 years) from minority populations in the United States had diminished cortisol reactivity compared to non-Hispanic whites (59). These results may not be transferable to children in other countries/cultures. As our results represent novel outcomes in childhood literature, they require replication in multiple samples.

5. Conclusion

Our results add novel evidence to the literature, suggesting that early childhood lead exposure is significantly associated with dysregulated HRV during an induced stress task in early adolescence, indicative of a dysregulated stress response. Additionally, children with BLLs > 5 mcg/dl had significantly greater LF/HF ratios, suggesting a more dysregulated stress responses, compared to children with BLLs ≤ 5 mcg/dl.

These findings have implications for future research regarding how early childhood lead exposure could affect neurocognitive, cardiovascular, and metabolic function via a dysregulated physiological stress response system. Dysregulated stress responses are associated with a host of health consequences including cardiovascular and metabolic diseases (60, 61) as well as impaired neurodevelopment and neurocognitive outcomes of general and higher order cognition (62). Additional work is needed to examine physiological stress responses as a mediator for the relationships between lead exposure and the multitude of neurocognitive, cardiovascular, and metabolic detrimental outcomes known to be associated with lead exposure. Such research is especially important in children whose brains are particularly sensitive to environmental insults and where chronic exposure may be most influential for autonomic functioning. Understanding these relationships could help to develop interventions to target this biological mechanism, such as reappraisal/mindset interventions (63), and thereby reduce the harmful effects of lead exposure for children at greatest risk. Future research is necessary to confirm this relationship at even lower levels of lead in children globally to best inform public health practices regarding lead exposure and health outcomes.