Is lead in tap water still a public health problem? An observational study in GlasgowBMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7063.979 (Published 19 October 1996) Cite this as: BMJ 1996;313:979
- Graham C M Watt, professor of general practicea,
- Andrew Britton, research and development coordinatorb,
- W Harper Gilmour, senior lecturer in medical statisticsa,
- Michael R Moore, reader in medicinea,
- Gordon D Murray, director, Robertson centre for biostatisticsa,
- Stuart J Robertson, operations scientistb,
- John Womersley, consultant in public health medicinec
- a University of Glasgow, Glasgow G20 7LR,
- b West of Scotland Water Authority, Glasgow G22 6NU,
- c Greater Glasgow Health Board, Glasgow G2 4JT
- Correspondence to: Professor Graham C M Watt, Department of General Practice, University of Glasgow, Woodside Health Centre, Glasgow G20 7LR.
- Accepted 18 September 1996
Objective: To assess the relation between tap water lead and maternal blood lead concentrations and assess the exposure of infants to lead in tap water in a water supply area subjected to maximal water treatment to reduce plumbosolvency.
Design: Postal questionnaire survey and collection of kettle water from a representative sample of mothers; blood and further water samples were collected in a random sample of households and households with raised water lead concentrations.
Setting: Loch Katrine water supply area, Glasgow.
Subjects: 1812 mothers with a live infant born between October 1991 and September 1992. Blood lead concentrations were measured in 342 mothers.
Main outcome measures: Mean geometric blood lead concentrations and the prevalence of raised tap water lead concentrations.
Results: 17% of households had water lead concentrations of 10 μg/l (48.3 nmol/l) or more in 1993 compared with 49% of households in 1981. Tap water lead remained the main correlate of raised maternal blood lead concentrations and accounted for 62% and 76% of cases of maternal blood lead concentrations above 5 and 10 μg/dl (0.24 and 0.48 μmol/l) respectively. The geometric mean maternal blood lead concentration was 3.65 μg/dl (0.18 μmol/l) in a random sample of mothers and 3.16 μg/dl (0.15 μmol/l) in mothers whose tap water lead concentrations were consistently below 2 μg/l (9.7 nmol/l). No mother in the study had a blood lead concentration above 25 μg/dl (1.21 μmol/l). An estimated 13% of infants were exposed via bottle feeds to tap water lead concentrations exceeding the World Health Organisation's guideline of 10 μg/l (48.3 nmol/l).
Conclusions: Tap water lead and maternal blood lead concentrations in the Loch Katrine water supply area have fallen substantially since the early 1980s. Maternal blood lead concentrations are well within limits currently considered safe for human health. Tap water lead is still a public health problem in relation to the lead exposure of bottle fed infants.
For a given tap water lead concentration mater- nal blood lead concentrations are much lower than they were in 1981
Tap water lead remains the main correlate of raised maternal blood lead concentrations
An estimated 13% of infants are exposed via bottle feeds to tap water lead concentrations of 10 μg/l (48.3 nmol/l) or more
Maternal blood lead concentrations are generally within limits considered safe for human health
Concern about the neurological toxicity of lead in young children1 2 3 4 5 6 has renewed interest in safety limits for lead in tap water.7 8 9 The current European standard for lead in drinking water is 50 μg/l (241.5 nmol/l), which may be reduced to the World Health Organisation's new guideline of 10 μg/l (48.3 nmol/l).9 We assessed the current lead exposure of infants (directly via tap water used in bottle feeds and indirectly via maternal blood) in a water supply area in which about half of households still have lead pipework10 and in which maximal treatment measures (namely, pH adjustment from 1978, orthophosphate treatment from 1989) have been introduced to reduce plumbosolvency.11 12 13 14 15 16
Subjects and methods
The study set out to measure tap water lead and blood lead concentrations in a random sample of 150 mothers (sufficient to estimate the geometric mean blood lead concentration with a standard error of 0.4 μg/dl (0.019 μmol/l) and to compare blood lead concentrations in groups of 80 mothers whose tap water lead concentrations were in the ranges <2, 2–9, 10–24, and 25–49 μg/l (<9.7, 9.7-43.5, 48.3-115.9, and 120.8-236.7 nmol/l (see table 1)), which gave 90% power to detect a ratio of mean blood lead concentrations of 1.4:1 at the 5% level of significance.
A total of 9243 women resident in the Loch Katrine water supply area gave birth to a live child between October 1991 and September 1992. To obtain sufficient numbers of households with raised tap water lead concentrations the study targeted all 1391 mothers living in areas in which water quality monitoring had shown a high prevalence of high water lead concentrations17 and a 30% random sample of mothers in the remainder of the water supply area.
All mothers were sent a postal questionnaire in 1993 and invited to return a 30 ml water sample from the household kettle. Home visits were carried out by a research nurse, who obtained a 4 ml sample of maternal venous blood after stasis, a repeat kettle water sample, and a daytime water sample, comprising a 1 litre sample taken from the drinking water tap with no prior flushing.
Blood lead determinations were carried out at Glasgow Royal Infirmary by means of graphite furnace atomic absorption spectrometry after blood deproteinisation18 19 in a laboratory participating in the United Kingdom National External Quality Assurance Scheme.
Water lead measurements were carried out at the National Measuring Accreditation Service accredited Strathclyde Water Chemistry Laboratory by atomic absorption spectrometry with electrothermal atomisation.20 External quality assurance procedures included participation in the Water Research Centre Aquacheck scheme.
As response rates varied with neighbourhood type (eight categories of postcode sector based on census based housing and household characteristics) population prevalence data were estimated by applying observed prevalences within neighbourhood types to the distribution of neighbourhood types in the target population.21
The study was approved by the Greater Glasgow Health Board's community and primary care research ethics committee.
A total of 1812 postal kettle samples were received (response rate 49%). Some 17% of all households and 43% (104/242) of households reporting the presence of lead pipework had kettle water lead concentrations of 10 μg/l (48.3 nmol/l) or more. A total of 83.5% of infants were wholly or partly bottle fed. Some 84.5% of households with bottle fed infants had tap water lead concentrations below 10 μg/l; 9.7% were in the range 10–24 μg/l, 4.1% in the range 25–49 μg/l, and 1.7% in the range 50 μg/l or more.
The research nurse obtained blood specimens from 342 mothers, including a random sample of 138 households and 204 other mothers stratified according to household kettle water lead concentration. A smaller than expected number of households with raised water lead concentrations (table 1) was due partly to the low prevalence and partly to the fact that daytime water samples (the legal standard) collected at the home visit provided a lower estimate of water lead exposure than kettle water samples.
The geometric mean blood lead concentration of the random sample of mothers was 3.65 μg/dl (0.18 μmol/l). In 86 households where lead concentrations were below 2 μg/l (9.7 nmol/l) in all three water samples the geometric mean blood lead concentration was 3.16 μg/dl (0.15 μmol/l).
There was a direct relation between tap water lead and maternal blood lead concentrations (table 1). The estimated proportions of cases of maternal blood lead concentrations above 5 and 10 μg/dl (0.24 and 0.48 μmol/l) which were attributable to a tap water lead concentration above 2 μg/l were 62% and 76% respectively (table 1).
The crude response rate to the postal survey of 49% was probably an underestimate of the true rate as 21% of respondents had changed address since the birth of their child on average 18 months previously. Probably some questionnaires were not received.
Tap water data collected in 1981 in association with a European Commission blood lead survey suggested that there had been a large fall in the proportion of households with tap water lead concentrations above 10 μg/l (48.3 nmol/l)—that is, from 49% of households in 1981 to an estimated 17% in 1993 (M R Moore and S J Robertson, personal communication, 1995).
The geometric mean blood lead concentration fell from 11.9 μg/dl (0.57 μmol/l) in 198122 to 3.7 μg/dl (0.18 μmol/l) in 1993 (fig 1). As blood lead estimations were carried out in the same laboratory with appropriate internal and external quality control procedures the trends are likely to be real and not confounded by measurement bias. Compared with previous studies22 the mean maternal blood lead concentration in 1993 was substantially lower in relation to a given tap water lead exposure (fig 1). The results were consistent with a reduction since 1981 of lead exposure from non-water sources such as food, air, and street dust.23 24 It is also possible that orthophosphate in the water supply reduced the bioavailability of lead in tap water.25
Tap water still accounts for about two thirds of cases of raised maternal blood lead concentrations. Non-water sources of lead exposure are the most likely explanation of the background geometric mean blood lead concentration of 3.16 μg/dl (0.15 μmol/l).
HEALTH IMPLICATIONS OF MATERNAL BLOOD LEAD
It is estimated that the effect of low dose exposure to lead is a two to three point decrement in intelligence quotient for a 10 μg/dl (0.48 μmol/l) increment in blood lead concentration in the range 10–20 μg/dl (0.48-0.97 μmol/l).6 26 As the mean maternal blood lead concentration in this study was below 5 μg/dl (0.24 μmol/l), with only 4% (5/138) of values above 10 μg/dl and no woman having a blood lead concentration above 25 μg/dl (1.21 μmol/l)5 27—even in the subgroup whose tap water lead concentrations were above 50 μg/l (241.5 nmol/l; table 1)—it may be concluded that current maternal blood lead concentrations in Glasgow are well within limits considered safe for adults.28
We could not estimate precisely the blood lead concentrations to which an unborn child might have been exposed, nor the lead exposure of infants fed with breast milk. However, the generally low mean maternal blood lead concentration, with over 96% of mothers having a blood lead concentration below 10 μg/dl (0.48 μmol/l), suggests that exposure via these routes is likely to be low. In general lead concentrations in breast milk are about one tenth of the concentrations in blood.29 The WHO recommends that when most children have blood lead concentrations below 10 μg/dl no further action is required.28
HEALTH IMPLICATIONS OF LEAD IN TAP WATER
The 1993 WHO guidelines for the quality of drinking water describe a guideline lead concentration of 10 μg/l (48.3 nmol/l) for water used to make up bottle feeds.10 In the Loch Katrine water supply area about 13% of infants are exposed via bottle feeds to tap water lead concentrations which exceed the WHO guideline.
A precautionary measure, therefore, might be to assess tap water lead in the homes of pregnant women and when concentrations are raised to advise the use of lead free bottled water during pregnancy or until lead pipework can be replaced. Mothers in these homes should be strongly advised in favour of breast feeding.
A fuller version of this report may be obtained by writing direct to GCMW. We thank Mary Stewart for the postal survey and home visits; Margaret Slowman for maintaining the research office; Dr Robert Low and colleagues for help with recruitment; Dr David Halls, of Glasgow Royal Infirmary, for supervising blood lead measurements; and colleagues at Strathclyde Water Chemistry Laboratory, Rutherglen, for analysis of water samples (Steven Davis and staff) and collating results (Jim Dillon). We also thank the mothers who took part and granted access to their homes. The views expressed are ours alone and do not necessarily represent those of the supporting organisations.
Funding The study was funded by the Chief Scientist Office of the Scottish Office Health Department and supported by Strathclyde Water Services (from 1 April 1996 the West of Scotland Water Authority) and the Greater Glasgow Health Board.
Conflict of interest Two of us (AB and SJR) are employed by the West of Scotland Water Authority.