Difference between revisions of "Health effects of lead in Europe"

 This page is a product of the EBoDE project. The final report of the EBoDE project has been published as a report in 2011[1] and also as web pages in Opasnet. These links lead to parts of the report. EBoDE project: main page | overview | contributors | data overview | Parma meeting | abbreviations | all pages Health effects in Europe: benzene | dioxins | formaldehyde | lead | ozone | particulate matter | radon | second-hand smoke | transport noise | environmental burden of disease | results by country

Lead is present in the environment due to former application of lead in gasoline, leaded drinking water pipes, and use of lead in paints and other housing materials. Exposures to lead originate from various sources including air, drinking water, food stuff as well as surfaces and consumer products.

Most of the health endpoints are significant at much higher exposure levels that are found in European population today. Exposure to lead has significantly decreased for many countries in the last two decades, especially since the phasing out of leaded gasoline and the replacement of leaded water pipes. For example, Figure 3-3 shows the reduction of internal exposure to lead in humans in German students between the 1980s and now (German Environmental Specimen Bank [Umweltprobenbank des Bundes], data available online at www.umweltprobenbank.de). Indeed, lead has been the success story in environmental policies, but the follow-up in exposure data in the general population is poor. [1]

FIGURE 3-4. Blood-Pb in German Students (1981–2009, geometric mean in μg/l, sampling location: city of Münster)

Selected health endpoints and exposure-response functions

The EBoDE project focuses on two endpoints that have been shown to be relevant at current exposure levels: mild mental retardation (due to IQ loss) and hypertensive disease (due to rise in systolic blood pressure). For the other health endpoints, i. a., no empirically sound exposure-response-relationships are available. Therefore, our results may underestimate the actual EBD of lead exposure in Europe. The extent of this underestimation cannot be quantified sufficiently.

The hypothesis of an effect threshold was rejected in several studies (Téllez-Rojo et al. 2006, Binns et al. 2007, Chiodo et al. 2004, Kordas et al. 2006). There is strong evidence for an association between B-Pb (blood lead) and negative effects on neuropsychological parameters at levels lower than 100 μg/l (Walkowiak et al., 1998; Canfield et al., 2004; Carta et al., 2005). Therefore, extending the dose-response curve to the range below 100 μg/l is possible. Lanphear et al. (2005) proposed a log-linear model for this curve.

Findings on lead’s effects on the central nervous system in the low-dose range are available from longitudinal and cross-sectional studies (Lanphear et al., 2005). These studies showed B-Pb and decrease in IQ points with B-Pb in children. The WHO model for IQ loss was recently updated to consider B-Pb levels above 24 μg/l. It has to be taken into account, however, that no threshold for mental retardation has been confirmed, yet. The exposure/response-function (ERF) in the WHO model is:

$IQloss = \frac{(B_{Pb} - 24)}{20}$ (Lanphear et al., 2005; see also Table 3-19 in section 3.12).

The population distribution of IQ is as defined as N(100;15). When the IQ falls below a diagnostic threshold, IQ loss is defined as mild mental retardation, which is the health endpoint used in this study. This threshold is set at 70 IQ points. We calculate the number of cases of mild mental retardation by estimating how many individuals in the target age group (children 0-4 years) exceed the diagnostic thresholds due to the lead exposure.

Several longitudinal studies have examined associations of blood pressure change or hypertension incidence in relation to lead concentration in blood or bone. Glenn et al. (2006) concluded that systolic blood pressure is associated both with acute changes in the blood lead level as well as with long-term cumulative exposure. Blood lead levels can increase in women over the menopause, as lead is released from bone. This may increase women’s risk of high blood pressure.

The current WHO model for increased systolic blood pressure in adults aged 20–79 years assumes a linear relationship between 50-200 μg/l (increase of 1.25 mmHg for males and 0.8 mmHg for females per increase of 50 μg/l B-Pb). Above 200 μg/l, an increase of 3.75 mmHg for males and 2.4 mmHg for females per increase of 50 μg/l B-Pb is assumed. The model does not account for aggravating effects of increased blood lead levels during the menopause.

The ERF for mean increase in the systolic blood (mmHg) in the WHO model is (B-Pb >50 μg/l) (Fewtrell et al, 2003):

$\Delta mmHg = \frac{(B_{Pb} - 50)}{40}$

The calculation of the numbers of cases of hypertensive disease is similar to the calculations for mild mental retardation. The population distribution of systolic blood pressure is defined as N(135, 15). When exposure exceeds the diagnostic threshold, of 140 mmHg, the increase in blood pressure is defined as hypertensive disease. We calculate how many individuals in the target age group (>15 year olds) exceed the diagnostic threshold due to the lead exposure. [1]

Exposure data

It is not easy to estimate lead exposure levels, because population exposure measurements are not regularly conducted, and because of the decreasing trends in lead concentrations which are not fully known. The most reliable way to account for all different possible exposure routes is to measure the body burden of lead. The commonly used exposure metric for such measurement is the blood lead level (B-Pb, whole blood, μg/l).

For the application of the WHO model for IQ loss, distributions of B-Pb (defined by percentiles) are necessary, stratified by specific age groups. This means that data are needed about different fractions of the population that are exposed to certain categories of B-Pb levels. No coherent international data sources were identified for lead. Hence, data from individual studies conducted in all participating countries were used. The year in which these studies were conducted differs between countries and in some cases the limited temporal coverage prohibited trend estimation. In these cases the most recent data have been used. It is clear that the limited temporal representativity of the lead exposure data poses a significant source of uncertainty. Due to well established lowering trends for lead this is expected to cause mainly unknown overestimation of exposures and effects.

The data are presented in Table 3-9 below and summarized in Table 3-21 in section 3.12. As shown in Table 3-9, lead data have been measured in different age groups in the different countries. Data from the German Environmental Survey (GerES) show that age is an important influencing factor for B-Pb levels in humans. As there is virtually no evidence for a significant reduction in B-Pb levels since the year 2000, the difference in age groups is assumed to be one of the most important sources of uncertainties when comparing the different countries. Unfortunately, B-Pb data are not sufficient to correct the country data for age.

TABLE 3-9. Lead data (μg/l) for different countries, measured in different age groups and years, used in the lognormal simulation to yield the required distributional parameters.
Country Estimates (2004) Age group Year
AM GM SD
Belgium 22 16 14-15 2000-06
France 26 18 18-74 2006-07
Germany 22 16 20-29 2004
Italy 39 24 18-64 2000
Netherlands 19 11 1-6 2005

AM: Arithmetic Mean; GM: Geometrical Mean; SD: Standard Deviation (estimated using coefficient of variation).

As indicated above, both of the exposure-response models used apply a threshold level (50 μg l-1 and 24 μg l-1). Therefore, it is necessary to assess the fraction of the population being exposed to levels higher than these threshold levels. A probabilistic simulation model was used to calculate the fraction of the population exceeding the threshold using mean and standard deviation data and assuming lognormal distributions. Standard deviations were estimated for the simulation using a coefficient of variation estimated from the Finnish data. [1]

TABLE 3-10. Population distributions of blood lead levels used in the simulation of threshold exeedances assuming log-normal distribution.
Country BE FI FR DE IT NL
Adults mean 22.0 16.0 25.0 22.0 39.0 19.0
SD 15.6 11.4 17.8 15.6 27.7 13.5
CV 0.71 0.71 0.71 0.71 0.71 0.71
Children mean 22.0 16.0 25.0 22.0 39.0 19.0
SD 15.6 11.4 17.8 15.6 27.7 13.5
CV 0.71 0.71 0.71 0.71 0.71 0.71

SD: Standard deviation, CV: coefficient of variation.

Uncertainties per stressor and comparison with other studies

A list of the most important sources of uncertainty for each stressor in the EBoDE calculations is provided in Table 5-1. Some of these are further explained below. In addition, we will compare our estimates to results of a selection of similar studies. Comparison of different studies on environmental burden of disease helps to understand the role of various methodological and strategic selections made in each study, like the selection of stressors or health endpoints.

Lead: The calculation focused on mild mental retardation and hypertensive disease only. WHO EBD estimates (Fewtrell et al., 2003) include cerebro-vascular and other cardiovascular diseases besides hypertensive disease; therefore the current estimates for lead are slightly lower than the WHO estimates. [1]

 Excluded health endpoints and related assumptions Exposure data Exposure response function Calculation method Level of overall uncertainty a) Likely over- or underestimation b) Lead Other cardiovascular diseases than hypertensive disease; kidney damage; miscarriages; other effects of the nervous system; declined fertility; alterations in growth and endocrine function; behavioural disruptions; hearing-threshold changes; hyperkinetic syndrome; lung and stomach cancers. MMR: proxy for all lost IQ points Differences in study year. Differences in studied age group. Incomplete data, temporal extrapolation and poorly known exposure trends Threshold level. Shape of ERF Evidence limited at prevailing low exposure levels. Estimation of threshold exceedances ** Underestimation due to excluded end-points

References

1. Otto Hänninen, Anne Knol: European Perspectives on Environmental Burden of Disease: Esimates for Nine Stressors in Six European Countries, Authors and National Institute for Health and Welfare (THL), Report 1/2011 [1] [2] Cite error: Invalid <ref> tag; name "EBoDe" defined multiple times with different content