Using physiologically-based pharmacokinetic modeling to assess the efficacy of glove materials in reducing internal doses and potential hazards of N-methylpyrrolidone during paint stripping

The study addresses how effectively different glove materials reduce internal exposure to N-methylpyrrolidone (NMP) during paint stripping, within the context of the U.S. EPA’s TSCA risk assessment. USEPA had identified several occupational scenarios with margins of exposure (MOE) below the target value of 30, indicating potential concern. While USEPA noted gloves can reduce exposure, it did not evaluate glove efficacy by material. This work refines the risk assessment by quantifying how glove material and its permeation characteristics influence internal dose metrics (Cmax for acute and AUC for chronic) and resultant MOEs, thereby informing whether appropriate PPE can mitigate risk without prohibiting NMP use.

A focused review identified key permeation studies for NMP across glove materials. Zellers and Sulewski reported no breakthrough for butyl-rubber gloves over 4 hours and permeation rates of 6–19 µg/cm²/min for other materials, with temperature dependence. Stull et al. evaluated multiple glove styles against NMP-containing and other stripping formulations, finding butyl rubber and laminate gloves most effective; permeation rates varied by nearly three orders of magnitude (<0.1–94 µg/cm²/min), and surrogate formulations did not always predict performance with commercial products. Crook and Simpson tested 20 glove types and observed that butyl rubber and laminate provided the greatest protection, polyethylene showed moderate permeation, and latex/nitrile exhibited high permeation with occasional acute failures; rates varied by more than two orders of magnitude (<0.1 to >34 µg/cm²/min) and varied by brand and formulation. These studies inform categorization of glove protection levels and parameterization of permeation in the PBPK model.

The assessment follows USEPA’s margin of exposure (MOE) framework, defined as MOE = IDTA/IDEA, where IDTA is the internal dose at the point of departure (POD) from the toxicity assessment and IDEA is the internal dose estimated for the exposure scenario via PBPK modeling. Acute risk uses peak blood concentration (Cmax, mg/L) and chronic risk uses area under the blood concentration–time curve (AUC, mg·h/L). Toxicity PODs were adopted unchanged from USEPA: BMDL01 for fetal resorptions based on Cmax = 216 mg/L for acute, and BMDL05 for decreased fetal body weight based on AUC = 411 mg·h/L for chronic. Exposure scenarios: From USEPA’s assessment, eight occupational scenarios with MOE <30 (various mid and high exposures for miscellaneous stripping and graffiti removal, with or without respirators) were selected. Inhalation and dermal vapor-through-skin pathways were retained as modeled by USEPA. The dermal liquid pathway was refined to incorporate glove permeation. Glove permeation data and categories: Literature-derived steady-state permeation rates (PR, µg/cm²/min) for NMP across glove materials were compiled. Permeability coefficients (Kp, cm/h) for glove materials were calculated using Kp = PR × CF / C, where C is the test concentration (mg/L) and CF = 0.001 mg/µg × 1000 cm³/L × 60 min/h. Based on permeation rates, gloves were grouped: minimal protection (>2 µg/cm²/min), moderate protection (1–2 µg/cm²/min), and maximal protection (≤0.3 µg/cm²/min). Representative Kp values (means and ranges) for each group were derived from the literature. Net dermal permeability with gloves: The PBPK model represented the hand as a multilayer barrier. Net permeability was computed via Kpnet = 1 / (1/Kpskin + 1/Kpglove), using USEPA skin permeability coefficients (Kpskin = 0.00205 cm/h for neat NMP; 0.000478 cm/h for NMP solutions) and glove Kp values from the literature. Assumptions included no significant accumulation of NMP between glove and skin and constant exposed glove surface area consistent with USEPA’s original scenarios. PBPK simulations: The human NMP PBPK model used by USEPA was applied without structural or parameter changes other than the refined Kpnet values for gloved conditions. For each of the eight occupational scenarios, simulations were run to estimate internal doses (Cmax for acute, AUC for chronic) with each glove category, with and without respirator use as specified. To isolate glove impact on the dermal liquid pathway, additional simulations excluded inhalation and dermal vapor pathways to compute glove protection factors (PF = IDno gloves / IDgloves) for each glove category. Resulting internal doses were used to compute refined MOEs relative to the unchanged toxicity PODs.

- Without gloves, USEPA identified eight occupational scenarios with MOE <30, with some as low as 0.1, indicating potential concern. - Incorporating glove permeation into PBPK simulations substantially increased MOEs, dependent on glove category: - Acute assessment: Using moderate-protection gloves yielded MOE ≥30 in half of the scenarios; using maximal-protection gloves yielded MOE ≥30 in all acute scenarios. - Chronic assessment: Using maximal-protection gloves yielded MOE ≥30 in all but one scenario. - Calculated glove protection factors (dermal-liquid-only PFs) spanned a wide range (approximately 1.1 to 1900), demonstrating strong dependence on glove material, NMP formulation, and internal dose metric (Cmax vs. AUC). - Literature data confirmed large variability in NMP permeation across glove materials (roughly two to three orders of magnitude), with butyl rubber and laminate gloves generally providing the greatest protection, polyethylene moderate, and latex/nitrile poor protection. - Toxicity PODs anchoring MOE calculations were BMDL01 (Cmax) = 216 mg/L for acute fetal resorptions and BMDL05 (AUC) = 411 mg·h/L for chronic fetal weight decrements, supporting internal-dose–based risk characterization.

The analysis demonstrates that selecting appropriate glove materials can reduce internal NMP doses sufficiently to meet or exceed the target MOE of 30 in most occupational scenarios previously identified as potentially hazardous without gloves. PBPK modeling enabled integration of glove-specific permeation into internal dose predictions for both acute (Cmax) and chronic (AUC) endpoints tied to developmental toxicity. Results indicate that maximal-protection gloves (e.g., butyl rubber or laminate) can generally reduce risk to acceptable levels, while moderate-protection gloves may be adequate for many acute tasks but less consistently for chronic exposure scenarios. These findings support risk management strategies centered on PPE guidance rather than outright prohibition of NMP in paint strippers, provided users select and correctly use highly protective gloves suited to NMP formulations. The work underscores the importance of material-specific glove guidance in product labeling and safety documentation to ensure adequate protection.

This study refines TSCA-related risk estimates for NMP in paint stripping by incorporating glove-material–specific permeation into PBPK-based internal dose and MOE calculations. It shows that high-performance gloves (butyl rubber and laminate) can render most acute and chronic occupational scenarios acceptable (MOE ≥30), while moderate-protection gloves provide partial mitigation primarily for acute scenarios. The results can inform practical risk-reduction measures such as specific PPE recommendations on product labels and safety data sheets as alternatives to bans. Future research should include: (1) expanded, standardized permeation testing across brands and realistic NMP formulations; (2) assessment of time-varying and intermittent contact, glove degradation, and temperature effects; (3) field validation of PBPK predictions with biomonitoring; and (4) evaluation of user compliance and fit/tear risks to refine real-world protection estimates.

- The refined assessment retained USEPA’s toxicity model choices and PBPK model structure; uncertainties in POD selection and model parameters carry forward. - Glove effects were modeled using steady-state permeation data and a multilayer barrier assumption with no accumulation of NMP beneath the glove; transient dynamics, occlusion effects, and pooling under gloves were not explicitly modeled. - Representative Kp values were derived from literature with variability across glove brands and formulations; categorization by protection level may not capture all brand-specific differences. - Nondetect permeation rates were handled as half the detection limit, potentially biasing Kp estimates for the most protective gloves. - Exposed glove surface area and task parameters were assumed consistent with USEPA scenarios; deviations in real-world use (tears, improper donning, extended wear) could alter protection. - Only eight occupational scenarios with initial MOE <30 were evaluated; generalizability to other tasks, consumer uses, or different environmental conditions may be limited.