Appendix 2.1 Landfill

1. Have studies been done on human populations?

 

Yes.

 

The literature search revealed more than 220 papers published about the hazards to health from landfill sites. Of these, 101 are primary studies about the health impacts of landfill sites and 23 about the health impacts of contaminated drinking water.  Six review papers were found which covered the epidemiological evidence linking health effects with landfill sites (Cantor, 1997, Johnson, 1997, Johnson, 1999, Miller, 1996, Sever, 1997,  Vrijheid, 2000). The drinking water studies were included in this section because an important source of exposure from landfill sites is leachate into groundwater. However, in many studies, the source of the contamination was not known. In some studies the source was leaking chemical storage tanks, in others, chemical accidents. Studies were not included if the water was contaminated by sewage (see section on sewage below). Only seven of the total are occupational health studies, the rest being studies about the health impacts on nearby communities.

The studies looked for links between the landfill sites and the following health outcomes: Reproductive outcomes/ developmental effects on children (31 studies), Cancer (29), Symptoms (28), Psychosocial impacts (19), Biomarkers (13), Health problems - not specified in abstract (14), Mortality (5), Injuries/poisoning (2)

2. Have hazards been identi f ied? Does the appearance of the hazard precede the health outcome? Is the association biologically plausible? Is there data on exposure?

 

No.

The main weakness of the studies about landfill health effects is the complete lack of exposure data. All use residence near the site as a proxy measure of exposure - i.e. data based on census tract, post code, or residence within 2 or 3 km of the site.  A few studies provided more detailed exposure data. For example, in a French study (Zmirou et al, 1994) individual exposure was estimated for one point in time, using a dispersion model of volatile air pollutants and the daily activity patterns of each individual within the area under investigation. The landfill site had been in operation for the previous 9 years. In this study, there were no statistically significant differences in consumption of prescription drugs.

Where the hazards from landfill sites have been identified, as is the case in the National Priorities List sites in the United States, it is possible to estimate exposure using the EPA Human Exposure Model (Wolfinger, 1989). The model is based on assumptions about the rate and toxicity of site emissions and can be used to estimate cancer risks from inhalation for each site in terms of risk to the maximally exposed individual (MEI risk), to the average individual (AEI risk), and to the population. The results of this type of analysis are uncertain and are based on risky assumptions. These remain estimates, not data. However, there is some biological plausibility in the association of congenital abnormalities and landfill sites due to the sensitvity of the fetus.

3. Are there ANY hypothesis-testing studies?

 

No.

Because of the lack of exposure data, the studies are hypothesis-generating studies rather than hypothesis-testing studies.

4. Have any of the hypothesis-testing studies controlled for possible confounding factors?

 

No.

With ecological studies of this type, it is impossible to control for other sources of pollutants. For example, the conclusion that the landfill site in Nant-y-Gwyddon may have been responsible for an increased rate of congenital abnormalities in residents near the site (Fielder et al, 2000) has been challenged by researchers who pointed out that a municipal incinerator operated in the same area just before the landfill site opened (Roberts et al, 2000). There was no direct evidence that the landfill, rather than the poorly performing and heavily polluting incinerator, was the cause of the adverse health outcomes. As well as other environmental pollutants from industrial and traffic pollution, there is usually concurrent exposure to occupational hazards, indoor air pollutants, tobacco smoke, alcohol, prescription drugs and recreational drugs.

 

5. Are there more than 20 hypothesis-testing studies consistently showing strong or moderate relative risks?

 

No.

There are more than 20 hypothesis-generating studies but the results were inconsistent, with some showing associations between landfill and various health impacts while other studies found no associations. Relative risks ranged from no association to strong.

In reviews, discussion papers, conferences and consensus meetings, many attempts have been made to determine whether the findings indicate real risks associated with exposure to landfill sites. There is general agreement with the cautious position taken at a meeting convened by the WHO Regional Office for Europe in 1998 which concluded:

“Many of the studies detected an increased risk of the studied diseases and symptoms in populations living close to the landfills. However, the evidence supporting the causality of the association is inconsistent and inconclusive. Probably the strongest suggestion for causality was generated by studies on reproductive outcomes, such as reduced birth weight or some birth defects. However, all studies lacked direct exposure assessment, and the limited sample size of most studies makes a more specific analysis impossible. … Considering all the uncertainties, the meeting concluded that the present data do add to a suspicion that population exposure to emissions from hazardous wastes may pose a risk to population health. The present studies are not powerful enough to indicate which of the characteristics of the very inhomogeneous group of landfills that are included in the studies might be responsible for the observed small increase in the risk.” (WHO meeting, 1998)

 

Judgement - insufficient