Estimation of the total daily oral intake of NDMA attributable to drinking water.

Fristachi A and Rice G. (2007) Journal of Water & Health, 5 (3); 341-55.

Changes in US EPS regulations on disinfection by-products are causing some United States drinking water suppliers to switch from free-chlorine to alternative disinfectants such as chloramine. Chloramination may reduce total DBP levels however it may result in the formation of nitrosamines such as N-Nitrosodimethylamine (NDMA). Some early studies have indicated that chloramination may lead to higher levels of NDMA than chlorination. NDMA been classified as a probable human carcinogen. This paper estimates NDMA concentrations in drinking water and food and calculates the average daily dose (ADD) for ingestion route exposures for the U.S. population. It also estimates the proportional oral intake (POI) of NDMA attributable to the ingestion of drinking water relative to that of NDMA present in food and formed endogenously in the human body. Three age groups were examined: bottle-fed infants (less than 6 months), children (6 months to 17 years) and adults (greater than or equal to 18 years). An exposure model was developed that used NDMA concentration point estimates for each source and parametric distributions of averaging times, exposure duration, intake rates and body weight under the assumption of independence of the parameters.

Food ingestion rates were developed for cereal, dairy, fish, meat, vegetables, beer, powered infant formula and foods high in nitrite and nitrate levels including beets, lettuce, spinach and pork. Drinking water ingestion rates were developed for adults and children and for infants (not accounting for water used to reconstitute powered formula) using mean and standard deviation values. NDMA concentrations in drinking water were estimated from a 2001-02 survey of 21 U.S. and Canadian drinking water treatment plants that reported a range of NDMA levels from below the minimum reporting level (MRL) of 6 x 10 -4 micro g/L to 2.3 x 10 -2 micro g/L. A point estimate for NDMA in reconstituted infant formula was calculated to be 8.3 x 10 -2 micro g/L. The distribution of NDMA concentrations in meat, fish, diary, cereal and vegetables were based on results of European studies conducted between 1987 and 1992. Distributions of NDMA concentrations in beer were developed from values reported in two surveys of North American beers. NDMA concentrations from endogenous formation from the nitrosation of ingested secondary and tertiary amines were estimated from an in vitro study resulting in an estimation of 0.37 micro g NDMA formed endogenously per gram of nitrate/nitrite-rich food.

Bottle-fed infants were found to receive their largest NDMA exposure doses from powdered infant formation reconstituted with drinking water contributing 0.07 micro g/day or 98% of total intake. The largest source of NDMA intake for children (9.96 micro g/day) and adults (23.1 micro g/day) is predicted to be from endogenous formation, contributing 99% of total intake for both age groups. The next largest source of daily NDMA intake for children and adults on average was from meat (0.04 micro g/day), contributing 0.30% of total intake. A sensitivity analysis was performed to measure the potential importance of the model inputs to the variance of the ADD estimates. Results indicated that endogenous nitrosation contributed over 99% to the variance and dominated the overall uncertainty and obscured the contributions of other model parameters. Therefore endogenously formed NDMA was excluded from the final sensitivity analysis.

The estimated Lifetime average daily dose (LADD) of NDMA in drinking water over a 75 year lifespan was estimated at 7.3 x 10 -7 micro/kg-day. This represented 0.02% to 0.003% of total LADD from all sources depending on whether in vitro or in vivo estimates of endogenous NDMA formation were used in models. Exogenous sources were found to contribute about 2.8% of total exposure, with the remainder due to endogenous production of NMDA. The results from this study suggest that the occurrence of NDMA in finished drinking water for the US population leads to low exposures relative to other ingestion route NDMA sources. The mean concentration of NDMA in drinking water would have to increase about 47-fold for the POI to reach 1% relative to all sources of NDMA included in the model studied here. The NDMA contribution from food is small however it is relatively high compared to NDMA in drinking water. The contribution of NDMA formed endogenously greatly increases the total NDMA encountered and reduces the contribution of contaminated drinking water ingestion to total NDMA exposure.

Comment This study considered only ingested NDMA and thus excluded cigarette smoke which is a significant exposure route in smokers and others exposed to cigarette smoke. The authors note that their estimates are based on limited data but are largely consistent with an independent estimate of NDMA intake performed recently by WHO. NDMA is only one of many nitrosamine compounds that may be present in drinking water or food or produced by endogenous formation.