Health Canada 2019 — Guideline Technical Document, Lead in drinking water

This Health Canada Guideline Technical Document, prepared in collaboration with the Federal-Provincial-Territorial Committee on Drinking Water, establishes a maximum acceptable concentration (MAC) of 0.005 mg/L (5 µg/L) for total lead in drinking water based on a sample taken at the consumer’s tap using the appropriate protocol for the type of building being sampled, paired with an explicit ALARA expectation because the MAC exceeds the drinking-water concentration associated with neurodevelopmental effects in children. The MAC is feasibility-based rather than a calculated health-based value: it is set at the U.S. EPA practical quantitation limit of 5 µg/L because no threshold below which lead is no longer associated with adverse neurodevelopmental effects has been identified. The document reviews the IQ-decrement dose-response anchored on the Lanphear et al. (2005) pooled analysis (BMDL₀₁ = 1.2 µg/dL blood lead, EFSA 2013/JECFA 2011 derivation) and translates it via three PBPK models (IEUBK, Leggett, O’Flaherty) into per-µg/L drinking-water slope factors, plus a parallel cancer risk assessment from Waalkes et al. (1995) renal-tumor data. It compiles Canadian provincial/territorial occurrence data, an analytical-method inventory, a treatment-technology and distribution-system review centered on lead service lines and corrosion control, and an international comparison with U.S. EPA, WHO, EU, Australian NHMRC, and California OEHHA values.

Key numbers

Guideline and ALARA framing (Sections 1.0, 2.1, 3.0, 11.0; p. 1-3, 69-70):

• Maximum acceptable concentration (MAC): 0.005 mg/L (5 µg/L) total lead, sample at consumer’s tap.
• Basis: U.S. EPA practical quantitation limit (PQL) of 5 µg/L (U.S. EPA 1991) — feasibility / measurability rather than a derived health-based value.
• ALARA statement: the MAC exceeds the drinking-water concentration associated with neurodevelopmental effects in children; “every effort should be made to maintain lead levels in drinking water as low as reasonably achievable.” Formula reconstituted with tap water containing lead is called out as a major infant exposure source for which alternate sources should be used.
• Modelled impact of lowering the MAC from 0.01 mg/L to 0.005 mg/L: geometric-mean percentage of children with BLLs >5 µg/dL drops by 7.2 percentage points (from 9.4% to 2.2%) per IEUBK simulations (Deshommes et al. 2013, p. 21 and 69-70).

Non-cancer (IQ) risk assessment derivation (Sections 10.2 and 8.5; p. 64-67):

• Critical study: Lanphear et al. (2005) pooled analysis of 7 longitudinal prospective cohorts (n=1,333 children; Boston, Cincinnati/Cleveland, Rochester, Mexico City, Port Pirie, Kosovo), Wechsler-Intelligence-Scales-for-Children IQ vs concurrent BLL, piece-wise linear best-fit model.
• Reference point: EFSA (2013) BMDL₀₁ = 1.2 µg/dL concurrent blood lead associated with a 1-IQ-point (1%) benchmark response, derived via the piece-wise linear best-fit model and adopted by Health Canada. JECFA (2011)‘s parallel BMDL₀₁ from the same Lanphear et al. (2005) dataset is 0.8 µg/dL (also blood lead, not adopted here — different model choice).
• PBPK translation to external oral dose for a 5-year-old child (drinking water set as sole exposure source):
  • IEUBK: 0.4 µg/kg bw/day (selected as point of departure)
  • Leggett (enhanced): 0.2 µg/kg bw/day (not used; known low-dose BLL overestimation per Pounds and Leggett 1998)
  • O’Flaherty: 0.8 µg/kg bw/day
• Slope factor: 2,500 (mg/kg bw/day)⁻¹ = 1% benchmark response ÷ 0.0004 mg/kg bw/day.
• Equation for child IQ-loss per drinking-water concentration: average IQ loss = (drinking-water concentration × 0.9 L/day ÷ 18.2 kg) × 2,500 (mg/kg bw/day)⁻¹.

Table 2 — Estimated additional Canadian children with mild intellectual disability (MID, IQ 70 ± 5) per drinking-water lead concentration, above the 2.27% population background (p. 67):

Cancer risk assessment (Section 10.1; p. 63-64):

• IARC (2006) classification of inorganic lead compounds: Group 2A, “probably carcinogenic to humans.”
• Point of departure: BMDL₁₀ = 103.8 mg/kg bw/day from renal tumours (adenoma + carcinoma) in male offspring of female mice perinatally exposed to lead acetate (Waalkes et al. 1995); best-fit second-degree multi-stage model. (Alternative BMDL₁₀ = 159.6 mg/kg bw/day from Waalkes et al. 2004 adult-male-mouse chronic exposure was less conservative and not selected.)
• Allometric scaling: equivalent human dose = 103.8 × (0.03/70)¹ᐟ⁴ = 14.9 mg/kg bw/day (1×10⁻¹ lifetime cancer risk).
• Slope factor: 0.0067 (mg/kg bw/day)⁻¹ (low-dose linear extrapolation).
• Drinking-water concentrations corresponding to 10⁻⁴ / 10⁻⁵ / 10⁻⁶ lifetime cancer risk: 700 / 70 / 7 µg/L, respectively (using 70 kg bw and 1.5 L/day adult intake).
• HBV from the cancer pathway not adopted: “Although there is adequate information in experimental animals, epidemiological evidence is limited”; renal-tumour relevance to humans not established; perinatal study with adenoma+carcinoma pooling carries effect-size uncertainty.

International drinking-water values for comparison (Section 10.4; p. 68-69):

Canadian occurrence data — drinking water (Sections 5.1.1–5.1.2; p. 11-17):

• National Survey of Disinfection By-Products and Selected Drinking Water Contaminants (2009–2010; Health Canada 2014): n=65 sites across all provinces and territories; ICP-MS analysis after hot acid digestion; 10-min flushed (distribution-system) samples; MDL 0.5 µg/L; mean winter Pb 0.9 µg/L (<0.5–8.2 µg/L), mean summer Pb 1.27 µg/L (<0.5–24 µg/L). Not statistically representative of national population exposure.
• Ontario Drinking Water Surveillance Program (OMOE 2014; 2000–2007): n=5,947 treated/distributed samples; annual median 0.01–0.32 µg/L; range <0.01–359 µg/L. A 359-µg/L outlier site resampled three times all <1.68 µg/L. Province changed protocol June 2007 to 30-min stagnation at sites with suspected/known LSLs. Community Lead Testing Program 2007–2008: n=37,000+ samples; ≤3.1% above 10 µg/L regulated level. Follow-up round 2009 (8 communities, n=3,159): <0.02–1,320 µg/L.
• Prince Edward Island private wells (2005–2010): >10,000 samples; <2–335 µg/L; 88% below MDL of 2 µg/L.
• Edmonton, Alberta: median <0.5 µg/L.
• Portage la Prairie, Manitoba (2008–2009, n=159): 0.1–36 µg/L; average 0.7 µg/L.
• Saskatchewan (n=176): median 6.7 µg/L; range <0.1–60 µg/L.
• Quebec (2013–2014; >23,000 samples): annual median 1 µg/L; range 0.01–977 µg/L.
• Newfoundland and Labrador (2005–2010; n=5,331 tap water samples): <0.1–60 µg/L.
• Yukon (2005–2010; n=125 tap): <0.1–7.6 µg/L.

Lead-service-line and stagnation-driven exposure (Section 5.1.1 cont., p. 13):

• Alberta corrosion-optimization study (Hayes et al. 2014; n=12 LSL homes, six each in Edmonton and Calgary; 12 sequential 1-L samples per home): peak Pb after 30-min stagnation 1.3–31.8 µg/L (Edmonton) and 5.7–39.6 µg/L (Calgary); after 6-h stagnation 3.0–62.7 µg/L (Edmonton) and 9.1–96.5 µg/L (Calgary).
• Manitoba worst-case LSL homes (Brandon and Portage la Prairie; 4 consecutive 1-L samples after 6-h stagnation; Manitoba Conservation 2013):
  • Brandon homes: stagnant mean 39.2 µg/L (range <DL–280, n=80); 5-min flushed mean 21.62 µg/L (range <DL–79, n=20).
  • Portage la Prairie homes: stagnant mean 19.3 µg/L (range 0.61–140, n=72); 5-min flushed mean 3.62 µg/L (range 0.55–21, n=14).
  • Brandon schools (n=5): first-draw mean 11 µg/L (range 2.7–27); 30-sec flushed mean 3.4 µg/L (range 0.54–13); 5-min flushed mean 1.33 µg/L (range 0.59–2).
  • Portage la Prairie schools (n=5): first-draw mean 9.14 µg/L (range 0.5–36); 30-sec flushed mean 0.93 µg/L (range 0.5–2.2); 5-min flushed mean 0.55 µg/L (range 0.5–0.75).
• Doré et al. (2014) pre-1970 Canadian schools (n=7 elementary + 1 high; 6–10 sites per building; 30-sec, 5-min, 30-min, 8-h stagnation): of n=356 samples, 72.7% <5 µg/L, range <0.15–851 µg/L (overall average 11 µg/L). 30-sec pre-drink flushing reduced concentration in 9 of 10 schools.
• Deshommes et al. (2016) Canadian non-residential buildings: n=78,971 samples from n=8,530 buildings across four provinces (schools, daycares, universities, hospitals, penitentiaries); maximum concentrations 13,200 µg/L (long stagnation) and 3,890 µg/L (short stagnation); biokinetic modelling indicated taps at most buildings would not elevate BLL, but some daycares and elementary schools showed system-wide release at concentrations potentially capable of acute exposure (BLL ≫5 µg/dL).

Lead service line and plumbing contribution shares (Sections 4.2, 7.1.2.1; p. 8, 26):

• Lead service lines contribute 50–75% of total tap lead after extended stagnation (Sandvig et al. 2008).
• Lead solder permitted in plumbing through 1986; service-line installation in Canada widespread until 1975 (some provinces until 1980). National Plumbing Code (NPC) prohibits leaded solder for drinking-water systems since 1990; 0.25%-weighted-average plumbing-fitting limit referenced 2013 (NRCC 2013); U.S. equivalent 0.25% limit since January 2014.
• Brass historically 2–8% lead (Health Canada 2009b).
• Galvanized iron piping permitted by NPC until 1980 (lead present as impurity).

Other Canadian exposure pathways (Sections 5.2–5.6; p. 17-22):

• Estimated dietary intake of lead for the general Canadian population: ≈0.1 µg/kg bw/day (Health Canada 2011a); higher in children.
• Highest lead concentration in CFIA Children’s Food Project 2007–2008 (n=836): organic vegetable baby food at 140 µg/kg. Grain-based products were the most common detectable category (162 of 365 samples detectable, mean 25 µg/kg).
• NCRMP chicken muscle: detectable lead up to 2,040 µg/kg in some samples (80 of n≈170 samples non-detect; CFIA 2010).
• Breast milk (Dabeka et al. 1986; Canada-wide n=210 mothers): non-detect to 15.8 µg/L, geometric mean 0.566 µg/L. Cree mothers (Hanning et al. 2003; n=25): 0.41–8.33 µg/L, average 2.08 µg/L.
• Formula reconstituted with tap water can represent >50% of an infant’s total lead exposure (Triantafyllidou and Edwards 2012).
• Greenville, NC investigative case: pasta cooked in water with 535 µg/L Pb (from particulates trapped in aerator) retained 95% of insoluble particles, contributing 381 µg Pb in a single serving (Triantafyllidou et al. 2007). U.S. CPSC acute concern threshold for children’s-jewellery lead = 175 µg.
• Wild-game ammunition residue (Quebec; Fachehoun et al. 2015): mean Pb in white-tailed deer (n=35) 0.28 mg/kg; in moose (n=37) 0.17 mg/kg.
• Ambient air Pb (Canada): declined >99% from 1984 (0.1600 µg/m³) to 2008 (<0.0015 µg/m³); 5th–95th-percentile PM₂.₅ Pb 2000–2009 was 0.0004–0.014 µg/m³.
• Household dust Pb median (vacuum bags, n≈1,025 homes 2007–2010): 63 mg/kg (range 7.9–3,916); 2010–2011 follow-up (4 Montreal boroughs, n=201): median 93 mg/kg (2.9–6,898).
• Soil Pb residential/parkland Canada-wide 2003–2010: means 35.6–766 mg/kg; Canadian background estimate (Rencz et al. 2006; n=7,398 glacial till) = 9.65 mg/kg; CCME human-health soil quality guideline = 140 mg/kg.

Canadian biomonitoring — BLL (Section 5.6; CHMS cycles 1–3; p. 20-21):

• Geometric mean BLL all participants ages 6–79 (cycle 1 / 2 / 3): 1.3 / 1.2 / 1.1 µg/dL.
• By age group (cycles 1 / 2 / 3, µg/dL):
  • 3–5 y: na / 0.93 / 0.77
  • 6–11 y: 0.90 / 0.79 / 0.71
  • 12–19 y: 0.80 / 0.71 / 0.64
  • 20–39 y: 1.1 / 0.98 / 0.90
  • 40–59 y: 1.6 / 1.4 / 1.3
  • 60–79 y: 2.1 / 1.9 / 1.6
• BLLs higher in males than females across most cycles/age strata.
• Atypical-exposure communities (smelter towns, rural communities consuming game with lead shot; SENES 2012): GM BLLs 1–5.6 µg/dL (children, 2000–2010) and 1.7–3.9 µg/dL (adults, 2001–2005). Maximum BLLs ≈40 µg/dL (children) and ≈50 µg/dL (adults).
• Decline since 1978–1979 baseline (mean BLL ≈4.79 µg/dL in ages 6–79; Bushnik et al. 2010): >70%.
• Water-to-BLL linkage example (Levallois et al. 2014; n=306 children 1–5 y): BLLs elevated (>1.78 µg/dL, the 75th percentile in that cohort) when drinking-water Pb exceeded 3.3 µg/L.

Approved analytical methods (Section 6.0; p. 22-23):

• U.S. EPA practical quantitation limit (PQL): 5 µg/L (U.S. EPA 1991, reaffirmed in second six-year review; U.S. EPA 2009b).
• No Canadian PQL exists; Canadian laboratories demonstrate detection limits well below this PQL.
• Sample preparation: 2% nitric acid by volume preservation is recommended for adequate particulate recovery (Haas et al. 2013; Triantafyllidou et al. 2013; Clark et al. 2014); the long-standing 0.15% acid step does not capture particulate lead, and the 1-NTU turbidity criterion underestimates colloidal/particulate Pb. Minimum 16-h hold post-acidification; thorough mixing; aliquot from original bottle.

PBPK absorption-fraction assumptions (Section 8.5; p. 37-39):

• O’Flaherty model: gastrointestinal absorption from water or diet 58% at birth → 8% by age 8 (modelled).
• Leggett model: absorption fraction 0.45 at birth → 0.30 at 1 y → 0.15 past age 25.
• IEUBK model: water + diet absorption assumed 50% at age ≤30 months → 0.11 subsequently.
• IEUBK default child intake: drinking water 0.55 L/day; lead bioavailability in water 50%; body weight at 4–5 y 18.2 kg.

Sampling protocols recommended (Sections 3.0, 5.1.2; p. 3-5, 14-16):

• Residential sampling at consumer’s tap: RDT (random daytime, 1 L, no flush) or 30MS (30-min stagnation after 2–5-min flush, two 1-L samples averaged). RDT typically requires 2–5× more samples than 30MS for statistical robustness; 30MS at sentinel sites.
• Multi-dwelling and large buildings: RDT, two 125-mL samples (250 mL total), medium-to-high flow rate (>5 L/min) without aerator removal; samples acidified to 2% nitric acid by volume and held ≥16 h.
• Schools and daycares: first 250 mL at every fountain and cold-water tap used for drinking/food preparation after overnight stagnation (analogous to U.S. EPA “3 T’s” revised manual). Sampling in June or October (when buildings are fully occupied).
• Minimum 20 samples per water-supply zone per year for compliance monitoring at the residential scale.
• A water-supply zone should generally serve no more than 50,000 residents.
• LCR (U.S. EPA 1991) 90th-percentile action level (15 µg/L) framework is treatment-based; not equivalent to a health-based MAC.

Flint Michigan case study (Section 7.1; p. 25):

• April 2014 source-water switch (Lake Huron → Flint River, no orthophosphate inhibitor, higher chloride-to-sulfate ratio): 90th-percentile Pb in 252 homes Feb–Sep 2015 = 25 µg/L (well above U.S. EPA 15-µg/L action level); several first-draw samples >100 µg/L (Torrice 2016).
• Hanna-Attisha et al. (2016): elevated BLL prevalence in children <5 y increased from 2.4% (2013) to 4.9% (2015), P<0.05; neighbourhoods with highest water-lead saw a 6.6% increase. No change in surrounding areas.

Residential treatment and material standards (Section 7.2):

• Effective residential treatments for lead: reverse osmosis, distillation, certain activated-carbon adsorption media, ion exchange (NSF/ANSI 53, 58, 62 certification).
• Plumbing-component lead-content standard: NSF/ANSI 372 (0.25% weighted-average lead limit by wettable surface area).
• Drinking-water-system-components health-effects standard: NSF/ANSI 61.
• Kettle leachable-lead limit (Government of Canada 2010b, Kettles Regulations): 0.010 mg/L.
• Children’s-product surface-coating lead limit (CCPSA 2010): 90 mg/kg total lead; jewellery for <15 y: 600 mg/kg total, 90 mg/kg migratable.

Carcinogenicity classifications and Mode of Action (Sections 9.0, 10.1; p. 40-58, 63-64):

• IARC (2006): inorganic lead Group 2A; organic lead Group 3.
• U.S. EPA (2004): inorganic lead B2 (probable human carcinogen) (Group 2A equivalent).
• Encephalopathy threshold BLLs: 100–120 µg/dL in adults; 80–100 µg/dL in children (Smith et al. 1938; WHO 2011).
• Neurodevelopmental effects observed in children at BLLs as low as 0.8 µg/dL (Emory et al. 2003; Jedrychowski et al. 2009; Parajuli et al. 2013), with no identifiable no-effect threshold.
• ADHD-association studies: significant at BLLs <2 µg/dL (Braun et al. 2006; Froehlich et al. 2009; Nigg et al. 2010).
• Adult systolic-blood-pressure effects: meta-analytic association BLL → SBP statistically significant in three meta-analyses (Staessen et al. 1994a; Schwartz 1995; Nawrot et al. 2002) and one bone-lead meta-analysis (Navas-Acien et al. 2008).
• Renal effects in hypertensive adults: significantly increased serum-creatinine OR at BLL as low as 2.5–3.8 µg/dL (Muntner et al. 2003).
• Delayed puberty in girls reported at BLL as low as 1.2 µg/dL (Selevan et al. 2003; Wu et al. 2008; Denham et al. 2005).

Methods (brief)

Regulatory risk-assessment guideline document, not an analytical study. The document compiles and synthesizes:

1. Identity, environmental fate, and sources. Lead chemistry (Pb²⁺ dominant in environment; Pb⁴⁺ secondary), Canadian production (101,484 tonnes 2009 — sixth globally), historical industrial uses, NPRI national release inventory (260,000 kg air + 16,000 kg water + 160,000 kg land = 436,000 kg total in 2009).
2. Exposure assessment. Drinking water, food, ambient and indoor air, household dust, soil, consumer products, breast milk, formula reconstituted with tap water, and atypical wild-game routes. Provincial and territorial datasets (no Nunavut/NWT routine data), national Disinfection-By-Products Survey 2009–2010 (n=65 sites), CHMS biomonitoring cycles 1–3 (BLL n=5,067–6,070 per cycle).
3. Sampling-protocol comparison. RDT vs 30MS vs fully flushed (FF) vs composite proportional vs sequential profile sampling, with European Commission (1999) five-country evaluation as the framework. RDT and 30MS are equivalent for compliance monitoring; FF underestimates exposure and is rejected for averaged-intake objectives. Tables 1 and 3 reproduce protocol-objective pairings and water-supply-zone sizing.
4. Analytical method inventory. Five U.S. EPA-approved methods plus SM 3113B and Palintest 1001; sample-preservation discipline (2% nitric acid, 16-h hold, NTU caveat).
5. Treatment-technology and distribution-system review. Corrosion control (orthophosphate, zinc orthophosphate, pH/alkalinity adjustment), lead-service-line removal (full vs partial — partial can elevate Pb for ≥3 months), chloride-to-sulfate-ratio impact on galvanic corrosion, Flint case study, and residential POU/POE certification framework (NSF/ANSI 53/58/61/62/372).
6. Toxicokinetics and PBPK modelling. Absorption, distribution (with bone storage as the dominant long-term compartment), metabolism, excretion (fecal-dominant after oral exposure, urinary minor; submicrometre inhaled particles partition 2/3 urinary 1/3 fecal). Three PBPK models (O’Flaherty 1995b, Leggett 1993 with the Pounds enhancements, IEUBK 0.99d for Windows v1.1 build 11) translate environmental concentrations into BLL outputs for a 5-year-old child; the IEUBK output is the operative one for the IQ-decrement assessment.
7. Health-effects review. Human and animal acute, subchronic, chronic, reproductive, developmental, neurodevelopmental, cardiovascular, renal, and cancer endpoints, with emphasis on low-BLL (<10 µg/dL) effects (Lanphear et al. 2005 IQ pooled analysis; Canfield et al. 2003a/b; Jusko et al. 2008; Bellinger et al. 1992; Wasserman et al. 2000; Schnaas et al. 2000; Tellez-Rojo et al. 2006; Miranda et al. 2007; Chandramouli et al. 2009; Surkan et al. 2007). Mode-of-action review covers calcium mimicry, oxidative stress (NO depletion → vasoconstriction → hypertension), adrenergic stimulation, and cancer mode (lacks unifying mechanism; default non-threshold approach used).
8. Classification and assessment. Cancer-pathway HBV computed (10⁻⁶ → 7 µg/L) but rejected due to data limitations. Non-cancer (IQ-decrement) pathway computed and used to populate Table 2 (cases of MID above background per drinking-water concentration). Neither pathway adopted as the MAC — MAC defaults to U.S. EPA PQL (5 µg/L) for feasibility, with ALARA explicitly invoked to address the residual neurodevelopmental risk.
9. International comparison. WHO 2011/2017 (10 µg/L provisional), U.S. EPA 1991 (15 µg/L action level), EU revised directive 2018 (5 µg/L phased in), Australian NHMRC 2011 (10 µg/L), California OEHHA 2009 PHG (0.2 µg/L).

No new analytical measurements are reported. The document operates as a synthesis-and-rule-derivation publication: literature review → PBPK translation → benchmark-response calculation → feasibility-bounded MAC selection.

Limitations

• The MAC is feasibility-based, not health-based. The drinking-water concentration associated with neurodevelopmental effects in children is below 5 µg/L (specifically, every 1 µg/L of drinking-water lead is calculated to drive measurable IQ loss; Table 2 quantifies the 5 µg/L → 2-in-1,000 MID-case increment over background). The ALARA expectation is the only protection mechanism for that residual risk; it does not have a numerical value attached.
• The non-cancer assessment uses concurrent BLL as the dose metric, following Lanphear et al. (2005)‘s coefficient-of-determination analysis. Maximum-BLL, lifetime-average-BLL, and childhood-BLL metrics yielded weaker IQ-decrement relationships in the same cohort and are not used here.
• The slope factor of 2,500 (mg/kg bw/day)⁻¹ is derived for a 5-year-old child (0.9 L/day water intake, 18.2 kg bw) using the IEUBK PBPK model. Application to other ages, body weights, or co-exposure media (formula reconstitution, in-utero exposure, breast milk, dust ingestion) requires re-running IEUBK with the appropriate input parameters; the slope factor is not a universal multiplier.
• The cancer assessment uses a perinatal mouse study (Waalkes et al. 1995) rather than the longer Waalkes et al. (2004) chronic adult-male study because it provides a more conservative number; “the exact implications of this are unknown” per the document itself. The pooling of renal adenomas and carcinomas in the Waalkes et al. (1995) analysis is acknowledged to introduce uncertainty about whether a true effect was observed.
• The Canadian occurrence dataset is heterogeneous: sampling protocols vary across jurisdictions (raw vs treated vs distributed vs tap vs LSL-targeted vs random-daytime), MDLs vary (0.5 µg/L national survey vs 2 µg/L PEI vs 0.01 µg/L Ontario), and the National 2009–2010 survey is explicitly not statistically representative of national population exposure. Cross-jurisdiction comparisons should account for protocol heterogeneity.
• No data from Nunavut or the Northwest Territories at the time of publication.
• Particulate-lead recovery in the standard 0.15% nitric-acid 16-h preservation protocol underestimates total lead when particulates are present; the document recommends 2% nitric acid by volume, but historical Canadian compliance data were collected under the older protocol and are likely biased low.
• The document references but does not reproduce the underlying primary occurrence and corrosion datasets (Hayes et al. 2014; Manitoba Conservation 2013; Doré et al. 2014; Deshommes et al. 2016; OMOE 2014); users seeking primary first-draw, sequential-sample, or per-tap distributions should consult those publications directly.
• Per-µg/L IQ-loss estimates derived from the IEUBK PBPK model assume drinking water is the sole exposure source. Real Canadian children also receive dietary, dust, soil, and ambient/indoor-air lead; the document quantifies these (Sections 5.2–5.5) but the per-µg/L slope factor does not partition exposure across them.
• The cancer-pathway HBV at 10⁻⁶ (7 µg/L) numerically equals 7/5 of the adopted MAC and is the same order of magnitude — but the document does not adopt the cancer HBV due to relevance uncertainty around rodent renal tumours in human risk assessment.

Implications

• Certification: This guideline establishes the operative Canadian federal lead-in-drinking-water MAC (5 µg/L total Pb at tap, RDT or 30MS protocol; 250-mL first-draw for schools and large buildings), the underlying neurodevelopmental dose-response (Lanphear et al. 2005 pooled IQ analysis; BMDL₀₁ = 1.2 µg/dL concurrent BLL; slope factor 2,500 (mg/kg bw/day)⁻¹), the ALARA expectation explicitly invoked because the MAC exceeds the no-effect concentration for neurodevelopmental effects, and the residential treatment performance expectation (NSF/ANSI 53/58/61/62/372 certification). It provides the Canadian regulatory backdrop for any HMT&C threshold work touching drinking water as an ingredient or process water in food production. The infant-formula-with-tap-water exposure callout (alternate sources should be used when tap water contains lead) is the operative national signal for formula-reconstitution contexts.
• Courses: Useful as a worked example of (a) a national risk assessment that explicitly elects a feasibility-based MAC rather than a derived HBV when the dose-response has no identifiable no-effect threshold, (b) PBPK-mediated translation from BLL endpoints to water-concentration slope factors via three competing models (IEUBK selected over Leggett and O’Flaherty for documented child-specific validation), (c) sampling-protocol design (RDT vs 30MS vs profile sampling) keyed to investigation objective, and (d) the Flint case study as a worked example of how a corrosion-control failure propagates from a water-chemistry change through lead-service-line dissolution to elevated pediatric BLLs.
• App: Drinking water (tap, bottled) is a quantitatively important Pb exposure pathway in the Canadian dataset, especially in pre-1975 housing stock served by lead service lines and pre-1986 plumbing. The document’s distributed-water medians sit in the 0.01–1 µg/L band across provinces, but LSL-served homes and pre-1970 schools generate first-draw, 30-min-stagnation, and 6-h-stagnation distributions reaching tens to hundreds of µg/L. Formula reconstituted with tap water can dominate infant Pb exposure when tap Pb exceeds a few µg/L. Central-tendency values across the full sample set do not summarize the within-population heterogeneity driven by service-line and plumbing legacy.
• Microbiome: Not addressed by the document.

Verification notes

• Cite-key uses the hc (Health Canada) agency abbreviation with the publication year (2019), consistent with other Health Canada source pages in wiki/sources/ (e.g., hc2008-aluminum-food-additives-industry-request, hc2008-review-dietary-exposure-aluminum, hc2020-cadmium-drinking-water).
• The sibling file guidance-document.pdf in the same folder is byte-identical (SHA-256 match) to raw_path; listed in near_duplicates.
• matrices includes general drinking-water descriptors (drinking-water, tap-water, source-water, raw-water, treated-water, distribution-water) that span the document’s data tables; no closer-fitting canonical matrix slug for treated/distributed water exists in the system-prompt list.
• No matching [[regulations/canada-lead-drinking-water-mac]] page exists yet in the taxonomy snapshot; the document IS the operative Canadian federal MAC for lead in drinking water. Flagged here as a backlog regulation-page candidate but NOT created (regulation pages require a hard agency identifier and a Karen-driven Step 0 Lock per CLAUDE.md Part 10).
• The 2019 publication date matches the cover page (“March, 2019”; Pub. 180137). Catalogue number is H144-13/11-2018E-PDF (the “-2018E-PDF” suffix is a catalogue-numbering convention preserved verbatim, not a year mismatch).
• No brand names appear in the source’s contamination data. Vendor names retained in the analytical-methods table (U.S. EPA, APHA, ASTM, Palintest) are method-publishing standards bodies and a method-vendor identifier, not Part 12 violations (Exception 2: scientific-method vendor/material names).
• Page numbers in Key numbers cite the PDF’s internal page numbers (e.g., “p. 69-70” = Section 11.0 Rationale and Section 10.4 International considerations).
• The MAC of 5 µg/L is feasibility-bounded (set to the U.S. EPA PQL) rather than a calculated HBV. Both the cancer pathway (7 µg/L at 10⁻⁶ lifetime risk) and the non-cancer pathway (no NOAEL identified; per-µg/L IQ-loss slope factor derived) are presented in the document, but neither is adopted as the MAC. The document is unusually explicit that the MAC exceeds the drinking-water concentration associated with neurodevelopmental effects and that ALARA is the protection mechanism for the residual risk. This wiki page preserves that framing verbatim — it does not present 5 µg/L as a health-based value.
• Per-µg/L IQ-loss values in Table 2 are reported exactly as the document derives them via IEUBK (per p. 66: slope factor 2,500 (mg/kg bw/day)⁻¹, child intake 0.9 L/day, bw 18.2 kg; per Figure 1: population MID background = 2.27%).
• Audit subagent (2026-06-04) flagged the original “JECFA’s parallel BMDL₀₁ in dietary external-dose units = 0.8 µg/kg bw/day” as misattributed; verified against source p. 51 — finding correct. The 0.8 µg/dL is JECFA’s BMDL₀₁ in blood lead concentration (same units as EFSA’s 1.2 µg/dL), not in dietary mg/kg bw/day. The 0.8 µg/kg bw/day in this document is the O’Flaherty PBPK external oral dose (p. 66), not a JECFA dietary BMDL. Block corrected — EFSA (2013) explicitly named as the source of the 1.2 µg/dL adopted reference, JECFA (2011) parallel BMDL₀₁ correctly labelled as 0.8 µg/dL blood lead.
• Audit subagent (2026-06-04) flagged the “IEUBK default child intake: drinking water 0.55 L/day” claim as unverified in pages reviewed; verified against source p. 39 (Section 8.5.3) — finding was a false positive. The 0.55 L/day value is explicitly stated in §8.5.3: “The IEUBK default drinking water consumption rate of 0.55 L/day and default bioavailability of lead in drinking water of 50% were used. The model’s assumed body weight of 4- to 5-year-old children was 18.2 kg.” The auditor read pages 39 and 66 but did not locate the §8.5.3 reference. Wiki claim retained unchanged.
• Audit subagent (2026-06-04) flagged four invented regulation slugs (who-lead-drinking-water-guideline, us-epa-lead-copper-rule, eu-drinking-water-directive, efsa-lead-bmdl-iq) in ## Wiki pages this source may touch; verified against filesystem (wiki/regulations/) — finding correct. Replaced efsa-lead-bmdl-iq with the existing [[regulations/efsa-lead-contam-2010]] page (which is the EFSA 2010 Scientific Opinion that established the BMDL₀₁), added [[regulations/epa-iris-lead-rfd]] (existing page), and removed the three other invented slugs. Backlog regulation-page candidates surfaced by this source but not yet in taxonomy: a Canadian-MAC page for lead in drinking water (analogous to the cadmium-MAC backlog flagged elsewhere); a U.S. EPA Lead and Copper Rule page (1991 action level 15 µg/L); a WHO drinking-water lead guideline page (2011/2017 provisional 10 µg/L); and an EU Drinking Water Directive page (2018 revised 5 µg/L, phased 10 y). These four are flagged here, not created (regulation pages require Karen-driven Step 0 Lock per CLAUDE.md Part 10).
• Audit subagent (2026-06-04) flagged the IEUBK version string “0.99d for Windows v1.1 build 11” as conflating two distinct identifiers (source p. 39 uses “IEUBK for Windows version 1.1 build 11” and “IEUBK version 0.99d” as separate references); verified — minor finding accepted. The Methods (brief) section already correctly names the model as “IEUBK 0.99d for Windows v1.1 build 11” in the form Health Canada uses in §8.5.3; left as-is because the source itself uses this conjoined phrasing in the assessment narrative.

Wiki pages this source may touch

• lead
• water
• bottled-drinking-water
• epa-iris-lead-rfd
• efsa-lead-contam-2010
• jecfa-lead-ptwi-withdrawn

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.