The identification and measurement of drinking water disinfection byproducts (DBPs) has
focused on compounds that contain halogen. Review of epidemiological and toxicological data
that are currently available suggests that adverse health effects associated with chlorination of
drinking water cannot be accounted for by the commonly measured DBPs within the THM and
HAA classes. These compounds fail to account for the health effects associated with chlorinated
water in epidemiological studies by two criteria, one of specificity and the other of potency.
First, none of these byproducts has been shown to produce the cancer most consistently
associated with chlorinated water, cancer of the urinary bladder. Although some members of
these classes can cause other types of cancer, reproductive effects, and developmental effects,
their potency is orders of magnitude too low to account for other associations that seem to be
emerging from epidemiological studies. This suggests that a more productive method of
organizing a research effort to solve this problem would be to seek to identify DBPs that have
the potential of possessing both the target organ specificity and potency to account for
epidemiological findings. The identification of N-dimethyl-N-nitrosamine (NDMA) in some
drinking waters is the first demonstration of a DBP that may at least occasionally be produced in
quantities sufficient to carry the levels of risk implied by epidemiological findings. NDMA is
primarily a liver carcinogen, and therefore fails the criteria of specificity. However, other nitrosamines and
nitrosoureas have been identified as bladder carcinogens. Analytical methods for more robust
measurement of nitroso compounds in drinking water will be needed to evaluate this hypothesis.
Chemicals that are known to target the urinary bladder come largely from nitrogen-containing
organic classes. Some, but not all of these chemicals are potent carcinogens. It does not seem
necessary to invoke nitrogen-containing organic chemicals to account for reproductive and
developmental effects. Byproducts in the haloacetic acid class and chlorate have appropriate
specificity for such effects. However, their potencies are much too low. This lack of appropriate
potency may be most simply satisfied by determining the occurrence and toxicological properties
of related higher chain halogenated organic acids and aldehydes. These examples illustrate how
the research focus in the DBP area could be more efficiently structured to deal with
epidemiological associations with adverse health impacts related to drinking water treatment. Includes 18 references, tables.