Shotyk and Krachler measured total lead in bottled waters from 125 brands across 28 countries. Excluding glass bottles, the median was 8.5 ng/L (n = 185), with a range from less than 1 to 761 ng/L. Twenty-five Canadian brands had a median of 15.9 ng/L (n = 25), range 2.1 to 268 ng/L. Pristine Ontario groundwater had a median of 5.1 ng/L (n = 18). This paper was also processed by Codex locally and is ingested here for cross-check.
Key numbers
The text layer prints an opening parenthesis where a plus-minus sign belongs, and a closing parenthesis where an equals sign belongs. The table below records the scientific reading. The ledger keeps the text-layer token.
Glass versus PET for three coded brands, ng/L lead:
| Code | Glass | PET | Ratio glass/PET |
|---|---|---|---|
| A | 227 ± 111 (n = 6) | 4.0 ± 1.6 (n = 5) | 56.8 |
| B | 261 ± 43 (n = 6) | 8.7 ± 3.3 (n = 6) | 30 |
| C | 154 ± 66 (n = 6) | 5.9 ± 3.7 (n = 6) | 26.1 |
The prose states that glass contained about 57, 30, and 26 times more lead than the matching PET. A footnote on bottle glass, measured by XRF, reports one bottle at 3 mg/kg lead and the others at 114 to 267 mg/kg. That is container glass, not the water. The detection limit for lead in water is given as 1 ng/L. SLRS-4 measured 88 ± 5 ng/L (n = 25) against a certified 86 ± 7 ng/L. The acid blank is about 800 pg/L, contributing about 0.8 pg/L to the sample.
The authors cite a maximum acceptable concentration of 10 µg/L used by the EU, Health Canada, and WHO. That cited limit is not a result of this survey.
Methods (brief)
Lead in water was measured by ICP-SMS after acidification. Container type is part of the result: glass, PET, and other packaging are not pooled in the headline median. The median of 8.5 ng/L excludes glass. Brand names are not used here. The three paired products are codes A, B, and C.
Evidence fitness
The measurements support total lead in bottled drinking water, with a strong container effect from glass. They do not support a mineral-water subclass, a brand ranking, or a lead-species split.
Limitations
The text layer substitutes parentheses for plus-minus and equals signs. Ratios in the table confirm the paired readings. The Canadian median is a subset, not the world median. The 10 µg/L figure is a cited regulatory maximum, three orders of magnitude above the medians in this paper, and it is not a measured concentration.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.