Waste management and public health: the state of the evidence
Appendix 2.4 Sewage
Bathing in sewage contaminated recreational waters and gastrointestinal symptoms
Because only a few studies investigated skin, eye, ear and respiratory illnesses associated with recreational use of contaminated water, this judgement is limited to the association with gastrointestinal symptoms. The judgement is based on a review paper by Pruss evaluating the health risks caused by poor microbiological quality of recreational natural water (Pruss, 1998). Water quality was measured by indicator-bacteria of faecal origin assumed to be resulting from sewage discharge. It is possible but unlikely that the contamination could be due to other bathers.
1. Have studies been done on human populations?
Yes.
Six review papers (Ashbolt, 1996, Barrell et al, 2000, IEH, 2000, Kindzierski and Gabos, 1996, Pruss, 1998) and 37 primary studies were found about the health effects of recreational bathing in sewage contaminated waters.
2. Have hazards been identified? Does the appearance of the hazard precede the health outcome? Is the association biologically plausible? Is there data on exposure?
Yes.
The hazards are microbial pathogens known to cause gastrointestinal symptoms. The exposure data consists of measurements of viral, bacterial and fungal pathogens and faecal indicator organisms typically found in sewage discharges.
3. Are there ANY hypothesis-testing studies?
Yes.
In the review by Pruss, there were 22 hypothesis-testing studies meeting strict criteria for inclusion (Pruss, 1998 p2)
4. Have any of the hypothesis-testing studies controlled for possible confounding factors?
Yes.
The confounding factors controlled for included food and drink intake, age, sex, history of certain diseases, drug use, personal contact, additional bathing, sun, and socioeconomic factors. 12/22 studies controlled for less than 3 of the previous factors. 4/22 studies took into account 3-4 factors. 6/22 studies accounted for 7 or more studies. Given the number of potential confounding factors, the pathogen threshold level for increased risk is still controversial. For example, it is possible that increased immunity in adult populations and in populations of countries with higher endemicity may result in higher threshold levels. Different countries detect different ranges of pathogens in water and use different detection methods.
5. Are there more than 20 hypothesis-testing studies consistently showing strong or moderate relative risks?
Yes.
Of the 22 studies in the Pruss review, 19 showed significant relationship of gastrointestinal symptoms to faecal indicator bacteria or bacterial pathogens. In 3 studies, there were no significant relationships. The relative risks included strong and moderately strong associations:
17 correlations where RR >2 (strong),
13 correlations where RR 1.5-2 (moderate)
18 corrrelations where RR <1.5 (weak)
6. Are there a range of study designs?
Yes.
There were 2 randomised controlled trials, 18 prospective cohort, 2 retrospective cohort studies.
7. Have studies been carried out in different population groups?
Yes.
Studies were carried out in the UK, USA, New Zealand, Hong Kong, Australia, Egypt, South Africa, Israel, Spain, France, Canada.
8. If dose-response relationships are observed, do they confirm the association between the hazard and the health outcome?
Yes.
Most of the studies showed significant dose-response relationship. The best dose-illness correlation was found with enterococci or faecal streptococci.
Occupational diseases of sewage treatment workers
1. Have studies been done on human populations?
Yes.
There was one review (Thorn and Kerekes, 2001)) and 38 primary studies. The health effects investigated were symptoms (17 studies), infections, i.e. hepatitis A, hepatitis C, legionella, leptospirosis, gastroenteritis (16 studies), mortality (3 studies), reproductive outcomes (1 study), biomarkers (3 studies) and cancer (5 studies).
2. Have hazards been identified?
Yes.
From the mortality and cancer studies, no hazards were identified. From studies on symptoms and infections, the following hazards were identified - bacteria, bacterial endotoxins, hydrogen sulphide, and organic solvents.
3. Does the appearance of the hazard precede the health outcome? Is the association biologically plausible?
Yes.
For symptoms, it is plausible that pathogenic micro-organisms, bacterial endotoxins, organic solvents and hydrogen sulfide could be related to the symptoms observed.
No.
For cancer, none of the agents commonly found in sewage treatment plants have been related to an increased risk of stomach cancer. The spread of the other cancers over a multitude of organs does not support a hypothesis of causality with agents commonly found in sewage treatment plants.
4. Is there data on exposure?
Yes.
Detailed exposure measurements were included in some of the studies on symptoms and infections but in most of the studies, the exposure was inferred by the subjects’ occupation as a sewage treatment worker. The exposure route was inhalation. Measurements were given of airborne viable bacteria (Lundholm and Rylander, 1983), (Melbostad et al, 1994), airborne endotoxin levels (Rylander, 1999),(Melbostad et al, 1994), hydrogen sulphide (Richardson, 1995), airborne organic solvents (Kuo et al, 1996), and amount of specific antibodies in the blood. For the mortality and cancer studies, no exposure data was provided.
5. Are there ANY hypothesis-testing studies?
Yes.
There were 29 hypothesis-testing studies. An example is a retrospective cohort study from the United States in which 28 sewage treatment workers were compared with data from a pooled non-exposed population (Kuo et al, 1996). The health outcome was central nervous system effects, determined by postural stability assessment. Exposure assessment was by measurement of organic solvents in the sewage treatment plant. In this, there was a statistically significant correlation between postural sway and organic solvent exposure and sewage workers had an increased postural sway compared with controls.
6. Have any of the hypothesis-testing studies controlled for possible confounding factors?
Yes.
Of the 29 studies, there were 16 which adjusted for personal factors such as smoking, alcohol use, age, educational level and gender.
7. Are there more than 20 hypothesis-testing studies consistently showing strong or moderate relative risks?
No.
There were 10 studies showing strong or moderately strong odds ratios (although there were no ORs in 4 of the studies).
8. Are there a range of study designs?
Yes.
Uncontrolled cohort, cross-sectional, case-control, case reports, and retrospective cohort studies.
9. Have studies been carried out in different population groups?
Yes.
Studies on sewage treatment workers in Germany, USA, Sweden, Denmark, Norway, UK, Canada, Greece, France, Israel, and Italy.
10. If dose-response relationships are observed, do they confirm the association between the hazard and the health outcome?
Not observed.
Judgement – probable.
Sewage discharges and reproductive outcomes
1. Have studies been done on human populations?
No.
Field and laboratory studies on a range of wild animals have demonstrated adverse reproductive outcomes from xeno-oestrogens, natural and synthetic substances with oestrogenic or anti-oestrogenic properties ((IEH, 1995)). These compounds occur in sewage discharges and have been associated with endocrine disruption in wildlife, including “thyroid dysfunction in birds and fish, decreased fertility in birds, fish, shellfish and mammals, gross birth deformities in birds, fish and turtles, metabolic abnormalities in birds, fish and mammals, behavioural abnormalities in birds, demasculinisation and feminisation of female fish and birds, and compromised immune systems in birds and mammals” (quoted in (IEH, 1995)). The relevance of these studies to human health is not clear but there is concern about the fall in quantity and/or quality of sperm in recent decades ((IEH, 1995; Colborn et al, 1997)).
Judgement – insufficient.
