Accounting for the impact of short-term variations in the levels of trihalomethane in drinking water on exposure assessment for epidemiological purposes. Part II: Biological aspects

Abstract

The variability of trihalomethane (THM) levels in drinking water raises the question of whether or not short-term variations (within-day) should be accounted for when assessing exposure to contaminants suspected of being carcinogenic and reprotoxic agents. The purpose of this study was to determine the magnitude of the impact on predicted biological levels of THMs (internal doses) exerted by within-day variations of THMs in drinking water. A database extracted from a campaign in the Québec City distribution system served to produce 81, 79 and 64 concentration profiles for the three most abundant THMs, namely chloroform (TCM), dichlorobromomethane (DCBM) and chlorodibromomethane (CDBM), respectively. Using a physiologically based toxicokinetic modeling approach, we simulated exposures (1.5 l water per day and a 10-min shower) based on each of these profiles and predicted, for 2000 individuals (Monte-Carlo simulations), maximum blood concentrations (Cmax), areas under the time versus blood concentrations curve (24 h-AUCcv) and total absorbed doses (ADs). Three different hypotheses were tested: [A] assuming a constant THM concentration in water (e.g., mean value of a day); [B] accounting for within-day variations in THM levels; and [C] a worst-case scenario assuming within-day variations and showering while THM levels were maximal. For each exposure profile, exposure indicator and individual, we calculated the ratios of values obtained according to each hypothesis (e.g., CmaxB/CmaxA and CmaxC/CmaxA) and the values corresponding to the 5th and 95th percentiles of these ratios. The closer these percentiles are to the value of 1, the smaller the error associated with assuming constant THM concentrations rather than their actual variability. Results showed that the minimal gap between these percentiles was TCM-AD(B)/TCM-AD(A) (5th=0.91; 95th=1.09), whereas the maximal gap was CDBM-Cmax(C)/CDBM-Cmax(A) (5th=0.50; 95th=3.40). Overall, TCM and ADs were the less affected (TCM<DCBM<CDBM and AD<AUCcv<Cmax) when accounting for within-day variations in water levels.