Overview
This cross-sectional study tested whether prevailing utility flushing guidelines (30 seconds to 2 minutes) meaningfully reduced lead concentrations in tap water at 376 New Orleans residences with high lead service line prevalence. Sequential 250 mL samples were collected at four flush durations: first draw (no flush), 30 to 45 seconds, 2.5 to 3 minutes, and 5.5 to 6 minutes. Flushing for 30 to 45 seconds did not reduce lead; median concentrations actually increased slightly relative to first draw. Statistically significant reductions were not observed until the 6-minute flush, and even at that point, 52 percent of sites retained detectable lead at or above 1 µg/L (the American Academy of Pediatrics school remediation threshold).
Key numbers
Median water lead levels by flush duration (µg/L, 376 to 218 sites per time point):
- First draw: 1.4 µg/L
- 30 to 45 s flush: 1.7 µg/L (median increase; p = 0.04 relative to first draw)
- 2.5 to 3 min flush: 1.4 µg/L (not significantly different from first draw)
- 5.5 to 6 min flush: 1.1 µg/L (statistically significant reduction vs. first draw; p < 0.05)
Maximum observed across all sample types: 58 µg/L. Detectable lead at or above 1 µg/L (LOQ): 60 percent of all samples across all flush durations. After 6-minute flush: 52 percent of samples still at or above 1 µg/L.
Fraction exceeding action level thresholds (across all sample types): US EPA Lead and Copper Rule action level 15 µg/L exceeded at some sites; WHO provisional guideline value of 10 µg/L exceeded at some sites; 88 percent of all samples were at or below 5 µg/L.
Older homes (pre-1950) and low-occupancy sites had significantly higher lead concentrations.
Methods (brief)
Sequential 250 mL samples collected from kitchen or bathroom faucet at four preset flush durations after 6 h or more stagnation. ICP-MS (Method 3125 B, Virginia Tech laboratory). LOQ: 1 µg/L. Non-detects assigned 0.5 µg/L for statistical analysis. City was in LCR compliance during the study period despite high lead service line prevalence.
Evidence fitness
Supports: demonstrating that 30-second to 2-minute flush guidelines are insufficient in cities with high lead service line prevalence; characterizing the distribution of residential water lead concentrations in a majority-LSL city; identifying older housing and low occupancy as risk predictors.
Does not support: generalizing flush efficacy to cities with lower lead service line prevalence or different distribution system materials; individual home risk assessment from a single visit; policy equivalence between this city and utilities without widespread LSLs.
Limitations
The study was conducted in a city where the utility was in LCR compliance but lead service line prevalence was 65 to 80 percent, an atypically high proportion that limits generalizability to systems with fewer LSLs. Flushing volumes tested are fixed by time; actual flushed volumes depended on flow rate, which varied by residence. The 6-minute flush was the terminal point; it is unknown whether longer flushing would achieve further reductions. Single-visit sampling does not capture within-home temporal variability.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.