Publications : 2026

Klaren WD, East A, Wheeler M, Vincent M, Wikoff DS, Covington T. Probabilistic estimation of daily total nitrosamine exposures generated from drinking water and dietary nitrate and nitrate sources. Abstract 3982, Society of Toxicology 65th Annual Meeting, San Diego, CA, March 2026.

Abstract

OPEN ACCESS

Background and Purpose: Research has suggested that exposure to nitrosamines can occur through endogenous generation via reaction between nitrate/nitrite and amines within the stomach, resulting in a comparatively larger exposure to total nitrosamines than to levels found in the diet and/or pharmaceuticals. Given the wide range of potential nitrate/nitrite exposure sources (e.g., drinking water, processed meats), information on the distributions of intakes, presence in food items, and constituent conversion rates are needed to understand the absolute and relative contributions of nitrosamine exposure estimates. The objective herein was to develop a probabilistic model to estimate population level exposures of daily total nitrosamines generated (DTNG) due to endogenous production from drinking water and dietary nitrate/nitrite sources. Methods: First, NHANES dietary survey data (2017-March 2020) based on dietary recall interviews was collected and reviewed for use within the model. Second, a targeted literature search was conducted to identify distributions of nitrate/nitrite, amine, and vitamin C content within various food groups, which were aligned to the food items reported in the NHANES dietary data. For each reported intake period in the survey data, the daily amount of nitrate/nitrite/amines/vitamin C was calculated as an input into the model. Additional literature searches were conducted to understand the range of nitrate-to-nitrite conversion rates within the saliva as well as the conversion of nitrite to nitrosamines (i.e., nitrosatability rate), the latter of which is dependent on the pH and the presence of potential inhibitors. Where possible, variables were incorporated into the model as distributions, or else accounted for via the pH change. Lastly, the model accounted for nitrosamine elimination upon exiting the stomach, where nitrosamine generation is inhibited by pH change in the small intestine. Based on available data, a range of N-nitrosodimethylamine (NDMA) half-life values was used as a surrogate for the total nitrosamine elimination distribution in the model. Ultimately, the model was run for each NHANES survey individual with random sampling from distributions for the variables indicated above. Since the survey included time of day for water and food ingestion, the model provides a minute-by-minute estimate of nitrosamine generation throughout the day for each survey participant and an estimate of the DTNG. Results: The probabilistic model estimated that the median DTNG for the simulated population was approximately 0.085 ug/day (IQR: 0.0159 – 0.397). The 95th percentile of estimated DTNG was 3.69 ug/day while the maximum DTNG was over 1,000 ug/day. The wide variability observed was driven by heavy tails of the distribution with much higher total daily nitrosamine generation. The drinking water and diet ingestion intervals (especially when intervals were <3-4hrs) impacted the amount of nitrosamines present at any given time due to elimination not occurring until after transiting the full stomach. A portion of the population did not ingest any drinking water during the day (~15%) or ingested drinking water without also ingesting an amine source resulting in no nitrosamine generation. Although the model indicates a wide variability in behaviors and generated nitrosamines across the population, a range of approximately two orders of magnitude was observed for most of the population. Conclusions: This model uses population dietary survey data and published information on food chemical constituents and nitrate/nitrite conversion relationships to estimate the DTNG that may occur following intake of nitrate/nitrite from drinking water and the diet. Notably, the model shows that 95% percent of the population generally falls within a relatively narrow range of DTNG and, while slightly lower, is within the range reported with previous models. Importantly, this more sophisticated model offers more granularity and accounts for the variability in a large number of the model parameters compared with previous estimates or in vitro analyses. Furthermore, future analyses that define the relative contributions of drinking water vs. dietary nitrate sources to the generation of nitrosamines may be useful for informing regulatory standards of nitrate/nitrite.