The relationship between air lead and blood lead in a modern US lead-acid battery facility: a longitudinal study

Abstract

1. Introduction

Lead is a human health hazard with a long and varied history of industrial uses. Occupational exposure to lead may occur in a variety of industries, including construction, lead mining, refining, sintering, smelting, and lead-acid battery manufacturing. Historically, the primary exposure route of concern has been inhalation, where deposition and retention of lead particles in the respiratory tract can subsequently be absorbed into the bloodstream. In addition to engineering controls and good hygiene practices, typical strategies to manage occupational inhalation exposures to lead in the workplace include monitoring concentrations of lead in the air (PbA) and mandatory respiratory protective equipment (RPE) when PbA concentration exceeds established thresholds.

It is widely accepted in the scientific literature that the concentration of lead in blood (blood lead, or PbB) is the exposure metric used to evaluate the potential effects of lead exposure on human health.¹ In adults, elevated PbB has been associated with adverse neurological, cardiovascular, renal, immunological, hematological, and reproductive health effects.¹ Workers' lead exposures are typically chronic in nature. And because lead accumulates in the bone, reducing lead exposures among workers is crucial in limiting their PbB levels, thereby protecting their health.^2,3

In 1978, the US Occupational Safety and Health Administration (OSHA) established a permissible exposure limit (PEL) for PbA concentration of 50 μg/m³ (averaged over an 8-hour period), which was largely determined by the PbA-PbB relationship characterized by a pharmacokinetic model.² The OSHA model was an adaptation of a model originally published by Bernard,⁴ later modified by the Center for Policy Alternatives to include consideration of job tenure and particle size.⁵ OSHA developed the model based on experimental data correlating lead particulate size to lead absorption (and subsequently affecting PbB levels), as well as a set of assumptions regarding lead particulate size distributions.² Several observational studies of the PbA-PbB relationship were considered at the time, and OSHA concluded that their results were similar to the estimates from the pharmacokinetic model.² Alongside the PEL, OSHA² also required measures, such as protective clothing, housekeeping (eg, vacuuming), filtered air lunchrooms, and hand and face washing prior to eating, to minimize the possibility of lead exposure via noninhalation routes and outside of the work environment (eg, ingestion of lead dust accumulated on surfaces, ingestion of lead-contaminated food).

In 2013, the California Office of Environmental Health Hazard Assessment (CalOEHHA) proposed an updated pharmacokinetic model (referred to as the “Leggett+” model) to predict the PbA-PbB relationship, aiming to improve the representation of lead physiology and better account for job tenure and particle size.³ Since its development, PbB concentrations predicted from the “Leggett+” model have been compared with those reported in a number of observational studies.^3,6,7 However, workplace PbA concentrations in these observational studies largely exceeded those typical of modern workplace environments and the OSHA PEL of 50 μg/m³; data at lower concentrations of PbA and PbB were sparse, limiting the reliability of reported PbA-PbB relationships. Uncertainty remains regarding how PbB concentrations predicted from the “Leggett+” model compare with those observed at lower PbA concentrations more typical of modern workplace environments.

In the present study, we examined the PbA-PbB relationship in a large longitudinal cohort of workers in a modern US lead-acid battery facility. Both PbA and PbB concentrations observed were relatively low, measured without RPE use, and representative of modern workplace environments with strict engineering controls and workplace regulations. Findings from this study can provide further observational evidence on the PbA-PbB relationship, as well as inform the performance of physiologically based lead models, at low PbA exposure concentrations. This study also improved upon prior observational studies by including a larger sample size, adjusting for a larger variety of potential confounders, and accounting for data correlations among repeated measurements within workers.

2. Methods

2.1. Study population

This study was conducted among workers in a modern lead-acid battery manufacturing facility in the United States, who worked in a lead environment (as determined by PbA data or the potential for lead exposure) and, while working in a lead environment, had periods of time with minimal to no respiratory protection requirements (based on company policy and historical employee PbA and/or PbB data). Upon careful review of accessible long-term employment records, a total of 236 eligible workers were identified and included in the study. The full employment histories (through November 2021) of all the included workers contained 4000 PbB measurements.

In general, a “pre-hire” PbB measurement was recorded at the beginning of each worker's employment history. During employment, a worker could change from a lead-exposed department/task to a non–lead-exposed department/task, or vice versa. Depending on the nature of the task being performed, the worker's familiarity with the task, the worker's most recent PbA and PbB levels, among other factors, he/she may or may not have been required to use RPE at the time of any given PbB measurement. Each worker's RPE requirement status at each PbB measurement was determined retrospectively by an industrial hygienist based on historical records.

In order to evaluate the association between PbA and PbB at lead-exposed departments and tasks without the effect of RPE use, we excluded PbB measurements that were taken at pre-hire (n = 216), at non–lead-exposed departments/tasks (n = 190), and when RPE use was required or when RPE requirement status could not be determined (n = 1147). After further excluding PbB measurements (n = 226) that had missing corresponding information for PbA measurement or covariates of interest (discussed in detail below), the final dataset for statistical analysis included 2221 PbA-PbB measurement pairs for the 236 workers. Ethical approval was not required, as all data were observational, deidentified, and extracted retrospectively.

2.2. Data collection

The dataset included personal PbA and PbB concentrations measured between May 2001 and November 2021. For any individual worker, PbA and PbB were generally measured on different days and at different frequencies, as the 2 measurements were taken separately for workplace regulation and safety purposes. To facilitate analysis, each PbB measurement of each worker was matched with his/her most recently available PbA measurement taken at the task, as a proxy for the concurrent PbA concentration.

A personal air sampling apparatus included a pump (AIRCHEK® 52; SKC, Eighty Four, PA, USA), a 3-piece cassette plastic filter holder, and mixed cellulose ester (MCE) membrane filters (0.8 μm, 37 mm) (SCK, or Millipore Sigma, Darmstadt, Germany). The filter holder was attached to the shoulder area of the worker, with a flow rate of 2.0 L/min for a minimum sample duration of 420 minutes over a single work shift. Air samples were collected in accordance with OSHA's Occupational Safety and Heath Standard for lead (29 CFR 1910.1025).⁸ Samples were analyzed on-site in an American Industrial Hygiene Association Laboratory Accreditation Programs (AIHA-LAP) accredited laboratory according to National Institute for Occupational Safety and Health (NIOSH) 7304/7303. Lead concentration was quantified by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) (iCAP™ PRO XP; ThermoFisher Scientific, Waltham, MA, USA) with a method detection limit of 0.015 μg/mL and reporting limit of 1.00 μg/mL. The results were reported as the 8-hour time-weighted average (TWA) for each worker and shift, converted to μg/m³ and rounded to integers.

PbB was collected and analyzed on-site in a Clinical Laboratory Improvement Amendments (CLIA) accredited laboratory. Prior to venous sample collection, the skin site was cleaned with diluted acetic acid and alcohol. Blood samples were diluted with water containing Triton X-100, and PbB concentrations were analyzed using Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (iCAP Q; ThermoFisher Scientific, Waltham, MA, USA). The method detection and reporting limits were 0.111 μg/dL and 1.00 μg/dL, respectively. Values were rounded to integers.

The department at which each individual worked at the time of each PbB measurement was recorded. Season of measurement was defined as winter (December, January, February), spring (March, April, May), summer (June, July, August), and fall (September, October, November). Job tenure (in years) for each worker at each PbB measurement was calculated by subtracting his/her date of hire from the measurement date. For 51 workers with nontypical employment histories (ie, 21 had rehire histories, 3 were originally hired as seasonal workers, and 27 were originally hired as temporary workers), job tenure was calculated from their original date of hire.

2.3. Statistical analysis

We first examined the characteristics of the included workers and their Pb measurements using descriptive statistics (ie, mean, SD, range, frequency, percentage). We examined the number of repeated measurements and the time interval between consecutive measurements for each worker, as well as the distribution of PbB measurements throughout each worker's employment history as indicated by his/her age and calendar time. The association between PbB and PbA was examined using a linear mixed-effects model with PbB concentration as the dependent variable, PbA concentration as a fixed effect, and worker as a random intercept to account for the correlations among measurements taken repeatedly over time within each worker. Both crude and adjusted models were performed. Variables adjusted for in the latter model included age, sex, job department, year of PbB measurement (ie, time trend), tenure (in years) at time of measurement, and seasonality (spring, summer, fall, winter) of measurement. In addition, a sensitivity analysis was performed further adjusting for pre-hire PbB, a potential source of extraneous variation for the dependent variable PbB at post-hire. Statistical analyses were performed using SAS software (Version 9.4; SAS Institute Inc, Cary, NC, USA) and R (v4.3.1; lme4, lmerTest, partR2, dplyr, merTools, ggplot2, and visreg packages).^9,‑16

3. Results

The 236 workers included in this analysis were from 9 departments, including Assembly, Cast on Strap (COS), Enveloping, Formation, Grid Casting, Janitorial, Pasting, Plate Preparation, and Quality Assurance. Activities conducted in each of these departments are described in Table 1. As shown in Table 2, the workers were mostly male (83%). The age at which a worker's first included measurement was made varied from 19 to 61 years (mean = 29.75 years). For 89% of the workers, the first included measurement was made in 2012 or after (ie, around the second half of the follow-up period). The job tenure at the time of a worker's last included measurement ranged from 1 to 24 years (mean = 6.04 years).

Figure 1

Blood lead (PbB) measurements by worker, age, and sample date. Data are shown for workers (n = 236) and PbB measurements (n = 2221) included in the adjusted model. Each line represents a distinct worker, and each dot on a line represents a PbB measurement for a worker with respect to calendar time and the worker's age.

Open in new tabDownload slide

Figure 2

Adjusted PbA-PbB relationship. The figure was generated from the linear mixed-effects model adjusted for age, sex, job department, year of PbB measurement, tenure at time of measurement, and seasonality of measurement. The dots represent PbA-PbB measurement pairs (n = 2221), with those in each unique shade representing repeated measurements from a unique worker (total n = 236 workers). The darker and narrower band represents the 95% confidence band, conditioning on male sex, an age of 31.85 years, job tenure of 3.85 years, and in the COS department in the spring of 2018. The lighter and wider band represents the 95% prediction band for any male worker who is 31.85 years old, has a job tenure of 3.85 years, and works in the COS department in the spring of 2018, considering only the variation in the fixed effect of PbA. COS, Cast on Strap; PbA, concentration of lead in air; PbB, concentration of lead in blood.

Open in new tabDownload slide

As for the PbA/PbB measurement pairs (n = 2221), 77% (1701/2221) were measured when workers were younger than 40 years, 82% were measurements of male workers, 88% were measured in 2012 or after, 62% were measured within workers' first 5 years of employment, and about 25% were measured in each season. The number of Pb measurements varied across departments, ranging from 2 in Janitorial to 923 in COS. More details about the distribution of Pb measurements throughout each worker's employment history as indicated by his/her age and calendar time are shown in Figure 1. The number of repeated PbB measurements per individual varied from 1 to 55, with about half of the workers having more than 6 repeated measurements (Figure S1). The time interval between 2 consecutive measurements for the same worker ranged from <1 to 155 months, with 6 months being the most common (see Figure S2).

Overall, the PbA concentration ranged from 1 to 50 μg/m³, with an arithmetic mean of 16.85 μg/m³; the PbB concentration ranged from 2 to 35 μg/dL, with an arithmetic mean of 15.47 μg/dL (Table 2). A crude comparison found a statistically significant difference in mean PbA concentrations across subgroups of sex, department, year, and tenure. Mean PbB concentrations also differed statistically significantly across those subgroups, as well as age and season of measurement. On average, female workers were exposed to higher PbA concentrations, but had lower PbB concentrations, as compared with male workers. The highest average PbA concentration was observed in the Plate Preparation department (18.32 μg/m³); the highest average PbB concentration was observed in the COS department (15.98 μg/dL). The lowest average PbA concentration was observed in the Janitorial department (5.50 μg/m³), although only 2 measurements were captured; the lowest average PbB concentration was observed in the Formation department, although only 3 measurements were captured. Both PbA and PbB concentrations were on average lower in 2012 or after, as compared with before 2012, and when measured at longer job tenure. For example, the average PbA and PbB were 17.46 μg/m³ and 16.14 μg/dL, respectively, for workers with less than 5 years job tenure, higher than the 14.71 μg/m³ and 13.72 μg/dL, respectively, for workers with 10 or more years of job tenure.

In the crude model, we observed a weak but statistically significant association between PbA and PbB. Specifically, a 1 μg/m³ increment in PbA was associated with a 0.081 (95% CI, 0.060-0.102) μg/dL higher PbB; the total variance in PbB that was explained by the fixed effect of PbA was very small (marginal |${R}^2$| = 0.021). This association was attenuated, but remained statistically significant after adjusting for age, sex, job department, year of PbB measurement, tenure at time of measurement, and seasonality of measurement (β = 0.025; 95% CI, 0.005-0.045 μg/dL); and the total variance in PbB explained by the fixed effect of PbA was also attenuated (marginal, semi-partial |${R}^2$| = 0.002).

The adjusted PbA-PbB relationship is visualized in Figure 2. Given the large sample size, we observed a fairly narrow 95% confidence band, conditioning on male sex, an age of 31.85 years, job tenure of 3.85 years, and in the COS department in the spring of 2018. However, consistent with the small |${R}^2$| estimate, we observed a wide 95% prediction band for any male worker who is 31.85 years old, has a job tenure of 3.85 years, and works in the COS department in the spring of 2018, considering only the variation in the fixed effect of PbA. When the variation in the random effect was further taken into consideration, the 95% prediction band covered almost the entire range of the observed PbB concentrations (figure not shown). Our result was robust to the additional adjustment for pre-hire PbB in the model (β = 0.026; 95% CI, 0.005-0.046 μg/dL; marginal, semi-partial |${R}^2$| = 0.002).

4. Discussion

The present study examined the PbA-PbB relationship in a large longitudinal cohort of workers in a modern US lead-acid battery facility, where the maximum PbA and PbB concentrations were 50 μg/m³ and 35 μg/dL, respectively. After adjusting for age, sex, job department, year of PbB measurement, tenure at time of measurement, and seasonality of measurement, we observed that a 1 μg/m³ increment in PbA was associated with a 0.025 (95% CI, 0.005-0.045; marginal, semi-partial |${R}^2$| = 0.002) μg/dL higher PbB.

The present study provides valuable evidence regarding the PbA-PbB relationship in an occupational setting at PbA and PbB concentrations that are lower than what have been observed historically. As shown in Table 3, PbA and PbB in most of the key observational studies to date ranged up to much higher concentrations, with only sparse data at relatively low concentrations. For example, Williams et al¹⁷ (see also Snee¹⁸) studied 39 battery workers in the United Kingdom (29 of whom provided complete information) whose highest reported PbA and PbB measurements were 298 μg/m³ and 93 μg/dL, respectively. Pierre et al¹⁹ studied 131 crystal manufacturing workers in France whose highest reported PbA and PbB measurements were 2131 μg/m³ and 61.3 μg/dL, respectively. Park and Paik²⁰ studied 117 workers in 4 different industries in South Korea whose highest reported PbA and PbB measurements were 7741 μg/m³ and 113.5 μg/dL, respectively.

Few studies focused on relatively low PbA and PbB concentrations only, and they were similarly limited by small sample sizes. Hodgkins et al²¹ studied 44 battery workers in the United States from 1983 to 1985, where 6-month average PbA and PbB concentrations ranged up to 33 μg/m³ and 40 μg/dL, respectively. The study reported that a 1 μg/m³ increment in cumulative PbA was associated with a 1.14 μg/dL higher PbB (P = .0003), after adjusting for paste machine job task, Black race, and smoking. Hodgkins et al²¹ noted that their PbA-PbB relationship finding was “strikingly higher” than those of extant studies in the battery industry where PbA concentrations were substantially higher, and offered as potential explanations for this difference, “non-linearity of the PbA-PbB curve, a higher fraction of large size particulate associated with higher PbA concentrations, survivor bias among workers exposed to higher PbA concentrations, and the cross sectional designs of most previous studies.” More recently, in a study of 32 battery workers in Japan from 2017 to 2020, Ono and Horiguchi²² reported a maximum worker PbA measurement (calculated as the geometric mean of 8-hour TWA PbA across all sample days for each worker) of 17.74 μg/m³ and a maximum PbB measurement of 18.0 μg/dL. This study reported that a 1 μg/m³ increment in PbA was associated with a 0.410 μg/dL higher PbB (P < .05), after adjusting for age, sex, and smoking.

Consistent with Hodgkins et al²¹ and Ono and Horiguchi,²² we observed a statistically significantly positive PbA-PbB association in the present study, albeit much smaller in magnitude. There are several possible explanations for the difference in magnitude. Primarily, the present study covered a much longer period of employment (up to nearly 20 years), as compared with the 2 prior studies (around 3 years), so our result represents a longer-term PbA-PbB relationship. The present study adjusted for a larger number of potential confounders, including time-related factors (eg, year, tenure, season), than the 2 prior studies, although we were unable to adjust for smoking like the 2 prior studies due to the lack of data for this factor. The 2 prior studies had much smaller sample sizes than the present study (Hodgkins et al²¹ followed only 44 battery workers; and Ono and Horiguchi²² followed only 32 battery workers); as a result, their point estimates were likely unstable (ie, with wide 95% CIs, although not reported explicitly). The present study addressed repeated measurements within individuals with the use of mixed-effects models, which fully used the information provided by how the measurements of each individual varied over time. In contrast, the 2 prior studies relied on averages of repeated measurements, thereby losing information about data variation during the averaging periods. In addition, there could be differences in particle size distribution between the different studies. We note, however, that our result is comparable to those of some other prior studies that examined largely higher PbA and PbB concentrations.^23,‑25

In the present study, we observed higher PbA (18.16 vs 16.56 μg/m³) but lower PbB (13.86 vs 15.82 μg/dL) on average among female workers than male workers. However, when allowing the estimated PbA-PbB relationship to differ by sex in our adjusted model, we did not detect statistically significant effect modification. It is worth noting that lead kinetics can differ between females and males.³ For example, the National Health and Nutrition Examination Survey (NHANES) 1999-2018 data (largely overlapping with our 2001-2021 study period) consistently showed lower PbB among females as compared with males in the general US population, a similar pattern to what we observed in the present study.^30,31 It is possible that our study did not have enough statistical power to detect effect modification by sex due to the small number of female workers (n = 39) in the study sample.

Although our analysis found a weak association between PbA exposures and PbB, we cannot exclude the possibility that our observed association was underestimated because we were unable to account for particle size (and the associated lead absorption rate), contribution to PbB by the exchange of lead in bone or tissue (eg, due to previous jobs) to blood, or potential PbA measurement error, which are common limitations of observational studies as previously identified by OSHA.² Regarding potential PbA measurement error, we note that we used each worker's most recently available PbA measurement taken at the job/task as a proxy for the PbA concentration concurrent to each of his/her PbB measurement. We were unable to assess how accurate this proxy was. Some consecutive PbB measurements were also matched with the same PbA measurement, when PbA was measured less frequently than PbB.

We also note that the present study is limited by the lack of data on each worker's potential lead exposure at the job from noninhalation routes (eg, dermal, ingestion) or outside the job (eg, diet, hobbies, home environment). Lead exposure through these non-PbA pathways, especially over a long period of time, could also contribute to the workers' measured PbB, as well as their lead storage in bones and tissues. We cannot exclude the possibility that the contribution to PbB from some of these non-PbA lead exposure pathways was greater than the measured PbA for some workers in the present study. More importantly, if workers at different PbA levels also had different levels of lead exposure from non-PbA pathways, our study results could have been confounded. For example, if workers with higher PbA exposure at the facility had lower lead exposure from non-PbA pathways combined, the true PbA-PbB relationship could have been masked because the observed PbB may not be higher at higher PbA levels. In addition, we did not have data on workers' body weight, which may affect lead kinetics and serve as a residual confounder in our observed PbA-PbB relationship.³

Because the present study only focused on workers with a particular type of employment history (ie, worked in a lead environment, and, while working in a lead environment, had periods of time with minimal to no respiratory protection requirements) and it is unknown if the included 236 workers were representative of all eligible workers at the facility during the study period, our results may not be generalizable to other workers. We used a worker's RPE requirement status as a proxy for his/her RPE use. As a result, workers who did not comply with their RPE requirement (eg, who did not wear RPE when exposed to PbA >50 μg/m³) were excluded from the study, and our results do not apply to such workers. Lastly, we note that our analysis might not have fully accounted for the procedures that were in place at the facility to protect the workers from lead exposure or from having elevated PbB concentrations. Such procedures could include job/task adjustment due to elevated PbB concentration, increased measurement frequency due to elevated PbA or PbB concentration, or use of other personal protective equipment.

According to company records, the workers in the present study had an average PbB level of 2.20 μg/dL at pre-hire, which is substantially lower than the observed average of 15.47 μg/dL at post-hire. The relatively low pre-hire PbB suggests workers' nonoccupational lead exposures were likely also relatively low, and that the observed increase in their PbB was most likely due to working at lead-exposed departments and tasks over time. If the PbA-PbB association is indeed weak, as our study and some previous observational studies have reported, the increase in workers' PbB levels (post-hire vs pre-hire) in our study may have been primarily driven by factors other than PbA alone (eg, hand-to-mouth transfer of lead deposited on surfaces). In that case, efforts to reduce occupational PbB may warrant consideration of measures beyond reducing PbA, such as measures that could affect alternative exposure routes. It is crucial that future observational studies quantify workers' lead exposures from noninhalation routes and nonoccupational sources to further understand the PbA-PbB relationship.

In the present study, we observed a very small |${R}^2$| statistic for PbA, indicating that the total variance in PbB that can be explained by the fixed effect of PbA was very small in our adjusted model. This was further demonstrated by the wide 95% prediction band showing that, at any particular PbA exposure concentration, a male worker who is 31.85 years old, has a job tenure of 3.85 years, and works in the COS department in the spring of 2018 could have a wide range of possible PbB levels based on model prediction. |${R}^2$| statistics for PbA specifically in the adjusted models were not reported in prior studies by Hodgkins et al²¹ or Ono and Horiguchi,²² although their reported model (as a whole) |${R}^2$| statistics (0.57 and 0.203, respectively) were comparable to that of the adjusted model in the present study (conditional |${R}^2$| = 0.564). We note that a similar model |${R}^2$| statistic (0.625) was reported by Lai et al,²⁷ a study that examined largely higher PbA and PbB concentrations. Although prediction was not the focus of the present study, the observed small |${R}^2$| estimate for PbA appears to indicate that a worker's PbA exposure concentration may provide less information about his/her PbB than predicted by pharmacokinetic models such as the “Leggett+” model, particularly in the lower end of PbA concentrations.^3,7

5. Conclusion

We present here a study investigating the relationship between PbA and PbB in a large longitudinal cohort of workers in a modern US lead-acid battery facility, where both PbA and PbB concentrations were relatively low. Our analysis appears to indicate that a worker's PbA exposure concentration may provide less information about his/her PbB than predicted by pharmacokinetic models such as CalOEHHA's “Leggett+” model. More importantly, we observed a weakly positive association of PbB with PbA that cannot seem to explain workers' PbB increase from pre-hire. Workers' PbB levels possibly have been primarily driven by their occupational lead exposures from noninhalational routes instead, in which case efforts to reduce occupational PbB may warrant consideration of measures beyond reducing PbA (eg, measures that could affect alternative exposure routes). Future observational studies with quantified lead exposures from noninhalational routes and nonoccupational sources are needed to examine this possibility and further understand the PbA-PbB relationship.

Acknowledgments

We thank Ms Charlotte Marsh and Mr Perry Piatos for their assistance with data preparation; Dr Ilkania Chowdhury-Paulino and Dr Leon Espira for their assistance with research quality control; Dr Teresa Bowers for her insights; and Mr Eric D. Fischbach and Mr Henry W. Adams for their administrative and editorial assistance, respectively.

Author contributions

The individual author contributions are as follows: Wenchao Li (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Validation, Visualization, Writing—original draft, Writing—review & editing), Jiayang Chien (Formal Analysis, Investigation, Validation, Writing—original draft, Writing—review & editing), Joel M. Cohen (Conceptualization, Funding acquisition, Supervision, Writing—original draft, Writing—review & editing).

Supplementary material

Supplementary material is available at Journal of Occupational Health online.

Funding

The International Lead Association (ILA) provided financial support to this work.

Conflicts of interest

All authors are employed by Gradient, an independent environmental and risk sciences consulting firm. The work reported in this article was conducted by the authors during their normal course of employment, with financial support by the ILA. The manuscript was solely conceived by the authors, and drafts were shared with members of the ILA. This article is the professional work product of the authors, and the opinions and conclusions within are not necessarily those of their employers or the financial sponsor of the work. Gradient has worked with several other organizations in the past that have an interest in lead science. None were involved with the conception or drafting of this manuscript.

Data availability

The data underlying this study were provided by a battery facility and are not publicly available due to confidentiality restrictions. Data can be shared on request to the corresponding author and upon obtaining permission from the facility.

References