The issue of lead exposure through drinking water was re-examined in light of modern public health goals, recent high-profile cases of elevated lead in water, and emerging concerns regarding the efficacy of legally mandated remedial strategies.

A critical literature review revealed that serious lead-in-water hazards are present at many US schools and homes, and that the threat to individuals is not eliminated by existing regulations. Health studies have provided strong links between lead in water and lead in blood of exposed populations, even at relatively low levels of exposure compared to reported lead occurrence in US tap water samples. As efforts shift from addressing pervasive lead sources that once elevated the blood lead of large percentages of the population, to more isolated individual cases requiring exceptional attention, the importance of carefully considering lead in water as a potential source for elevated blood lead increases.

Consistent with decades of prior research linking elevated water lead to elevated blood lead (EBL), lead-contaminated water in the high-profile case of Washington DC markedly increased the incidence of EBL for very young children. Specifically, incidence of EBL for children aged ≤ 1.3 years increased more than 4 times during 2001-2003 when lead in water was high, compared to 2000 when lead in water was low. The incidence of EBL for children aged ≤ 1.3 years was highly correlated (R² = 0.81) to 90th percentile lead-in-water levels from 2000-2007, and the risk of exposure to high water lead levels varied markedly in different neighborhoods of the city. Analysis conducted herein focused on identifying "worst-case" neighborhoods and populations. Specifically, this was the first study of the Washington DC case to focus on infants who are most vulnerable to harm from lead in water, and to perform smaller area analysis at the neighborhood (i.e., zip code) level in order to capture pockets of high risk among local communities.

Prior biokinetic modeling efforts, examining the potential adverse impacts of lead-in-water exposure, were re-examined to explicitly consider new public health goals. This included impacts on the most sensitive population groups (e.g., young children and particularly formula-fed infants), the potential variability in blood lead levels (BLLs) amongst exposed individuals within those groups (e.g., most sensitive children at the upper tail of the BLL distribution), more conservative BLL thresholds reflecting low-level adverse effects (e.g., 5, 2 and 1 µg/dL versus 10 µg/dL), and the possibility of acute health impacts. This re-evaluation creates a paradigm shift, in that levels of lead in water that were previously considered inconsequential are demonstrated to be of concern in specific circumstances.

The replacement of lead service lines in front of consumers' homes is a costly, federally mandated remedial action if a water utility exceeds the US EPA lead action level. Because utilities do not own the entire lead service line, they often only replace the portion of the service line up to the property line, typically with copper pipe. Experiences in Washington DC, as revealed by Freedom of Information Act requests, indicated that partial pipe replacements were not decreasing lead in water, and were actually associated with relatively high incidence of childhood lead poisoning. This prompted the first comprehensive investigation of potential long-term problems arising from galvanic corrosion between the remaining lead pipe and the newly installed copper pipe. Bench-scale experiments demonstrated that galvanic connections between lead pipe (new or aged) and copper pipe increased lead release into the water by 1.1-16 times, when compared to a full length of lead pipe alone. The small area of lead pipe adjacent to the copper joint (<0.5 ft) was gravely affected by galvanic corrosion, and accumulated a thick lead-rust layer (1 inch wide) that constituted a reservoir for semi-random particulate lead detachment into the water.

The work on simulated partial pipe replacements revealed that under worst-case scenarios of highly contaminated water samples, most of the lead was not quantified if water samples were not mixed thoroughly after standard preservation (i.e., after addition of 0.15% v/v HNO₃), or if water samples were transferred from one bottle to another prior to preservation. While there is no reason to believe that sample handling and pre-treatment dramatically skew regulatory compliance with the US EPA lead action level, slight variations from one approved protocol to another may cause lead-in-water health risks to be dramatically underestimated. This is of special concern in unusual situations of "worst-case" individual exposures to highly contaminated water, associated with childhood lead poisoning.

This work provides the water industry and health agencies with important new insights and perspectives on an old problem. Results can improve strategies to detect and mitigate lead-in-water hazards for individuals or populations, and inform future revisions to the US EPA Lead and Copper Rule.