Case Reports Giving Evidence of Short-Term Chealth Risks from Acute Exposure through Drinking-Water
Some case studies provide evidence that exposure to cyanobacterial toxins in drinking-water can lead to illness and even death. Due to the inability to identify the toxins at the time, the earliest reported cases offer only circumstantial evidence of a link between exposure to cyanotoxins and human illness.
- • Gastroenteritis associated with cyanobacteria was observed in the population of a series of towns along the Ohio River in 1931. Low rainfall had allowed the water of a side branch of the river to develop a cyanobacterial bloom which was then washed into the main river. As this water moved downstream, a series of outbreaks of illness were reported (Tisdale, 1931).
- • In Harare, Zimbabwe, children living in an area of the city supplied from a particular water reservoir developed gastroenteritis each year at the time when a natural bloom of Microcystis was decaying in the reservoir. Other children in the city with different water supplies were not affected (Zilberg, 1966).
- • In an incident in Sewickley, Pennsylvania, 62% of the population connected to a filtered, chlorinated drinking-water supply developed symptoms of gastroenteritis within a period of five days. The water, sourced from groundwater contaminated by an intrusion from the Ohio River, was treated and then held in open holding reservoirs prior to distribution. One reservoir had over 100 000 cells/mL of Schizothrix calcola, Plectonema, Phormidium and Lyngbya in the open water. The reservoir had just been treated with copper sulphate when the poisoning event occurred (Lippy & Erb, 1976). Although not known to be toxic at the time, Schizothrix, Phormidium and Lyngbya have all since been shown to be toxin producers elsewhere (Falconer, 2005).
While these reports note that the health effects could not be attributed to infectious agents, a caveat on this conclusion is that many of the aetiologic agents leading to the described symptoms were unknown at the time (e.g., viruses) or not detectable with sufficient sensitivity by a standard laboratory (Giardia, Cryptosporidium). The following later study addressed many of these issues.
• An outbreak, with a high death rate attributed to cyanobacterial toxins in drinking-water, occurred in the Paulo Alfonso region of Bahia State in Brazil following the flooding of the newly constructed Itaparica Dam reservoir in 1988. Some 2000 gastroenteritis cases were reported over a 42-day period, and 88 deaths, mostly children, occurred (Teixera et al., 1993). Blood and faecal specimens from gastroenteritis patients were subjected to bacteriological, virological and toxicological testing, and drinking-water samples were examined for microorganisms and heavy metals. No infectious agent was identified, and cases occurred in patients who had been drinking only boiled water. The cases were restricted to areas supplied with drinking-water from the dam. Clinical data and water sample tests were reviewed, and it was concluded that the source of the outbreak was water from the dam and that a toxin produced by cyanobacteria (Anabaena and Microcystis in high densities) was the most likely responsible agent, although the toxin could not be identified.
A closer association between human illness and exposure to cyanotoxins is demonstrated when the cyanobacteria were shown to be toxin producers, as illustrated in the following examples:
- • In Armidale, Australia, the water supply reservoir had been monitored for blooms of toxic Microcystis for several years, and MC-YM had been identified in these blooms. When a particularly dense bloom occurred, the water supply authority treated the reservoir with 1 mg/L of copper sulphate, which lysed the bloom, possibly causing a pulse of toxin release from the cells. An epidemiological study of the local population indicated subclinical liver damage occurring simultaneously with this treatment of the bloom (see Box 5.2).
- • A more severe outbreak of cyanobacterial toxicity in a human population occurred on Palm Island, off the north-eastern coast of Australia in 1979. Complaints of bad taste and odour in the water supply were attributed to a cyanobacterial bloom, and the authorities therefore treated the reservoir with copper sulphate. Within a week, numerous children developed severe hepatoenteritis, and a total of 140 children and 10 adults required hospital treatment (Byth, 1980). A CYN- producing strain of Rapbidiopsis raciborskii was later identified as the agent most likely to be responsible for this episode (see Box 5.3).
BOX 5.2: TOXIC MICROCYSTIS IN THE ARMIDALE WATER SUPPLY RESERVOIR AND PUBLIC HEALTH
At the time of this study, the city of Armidale, New South Wales, Australia, had a drinking-water supply from a eutrophic reservoir which had been experiencing repeated blooms of cyanobacteria since the early 1970s.
In 1981, a particularly extensive toxic bloom of Microcystis aeruginosa was monitored during its development. During the bloom, complaints of bad taste and odour in the drinking-water were received, leading to copper sulphate treatment of the reservoir. The toxicity of the bloom was monitored by mouse bioassay. A toxin had previously been isolated from Malpas Dam and partially described, which was later characterised as MC-YM (Botes et al„ I98S). This event was used as the basis for a retrospective epidemiological study of liver function in the population consuming the water, compared with a population in the same region supplied from other reservoirs. The data for the activity of plasma enzymes reflecting liver function were obtained for patients having blood samples examined at the Regional Pathology Laboratory for the 5 weeks prior to the bloom, the 5 weeks of peak bloom and its termination and for 5weeks after that. The data were then separated into those from patients having used the Malpas drinking-water supply and those using other supplies.
Serum enzymes reflecting liver function in patients consuming drinking-water from Malpas Dam or from other supplies included GGT=y-glutamyl transferase; ALT=alanine aminotransferase; AST=aspartate aminotransferase and AP=alkaline phosphatase (Falconer et al., I983). As shown in the figure above (redrawn from Falconer et al., 1983), y-glutamyl transferase in the blood of the group using the Malpas Dam water supply during the peak of the bloom and its lysis with copper sulphate was significantly higher than that in the same population before and after the bloom, and higher than that in the other population served by different water supplies. The clinical record gave no evidence of an infectious hepatitis outbreak or disproportionate alcoholism (Falconer et al., 1983). While the mean increase in y-glutamyl transferase activity was indicative of minor liver toxicity, some individuals within the population studied showed highly elevated enzyme activity, indicating substantial liver damage. This enzyme has also been shown to be elevated as a result of Microcystis toxicity in experimental studies with pigs and rodents, where it is used as an effective marker for liver injury (Fawell et al., 1993; Falconer et al., 1994).
BOX 5.3: PALM ISLAND MYSTERY DISEASE
In 1979, there was a major outbreak of hepatoenteritis among the children of an Aboriginal community living on a tropical island off the coast of Queensland, Australia. Altogether 140 children and 10 adults required treatment, which was provided by the local hospital for less severe cases and by the regional hospital on the mainland for severe cases possibly requiring intensive care. Diagnostic information included a detailed clinical examination showing malaise, anorexia, vomiting, headache, painful liver enlargement, initial constipation followed by bloody diarrhoea and varying levels of severity of dehydration. Urine analysis showed electrolyte loss together with glucose, ketones, protein and blood in the urine, demonstrating extensive kidney damage. This was the major life-threatening element of the poisoning. Blood analysis showed elevated serum liver enzymes in some children, indicating liver damage. Sixty-nine percent of patients required intravenous electrolyte therapy and, in the more severe cases, the individuals went into hypovolaemic/acidotic shock. After appropriate treatment, all the patients recovered (Byth, 1980).
Examination of faecal samples and foods eliminated a range of infectious organisms and toxins as possible causes for the outbreak and failed to identify the cause, hence the name “Palm Island Mystery Disease”. The affected population, however, all received their drinking-water supply from one source, Solomon Dam. Families on alternative water supplies on the island were not affected by the disease. Prior to the outbreak of the illness, a bloom of cyanobacteria occurred in Solomon Dam. The bloom discoloured the water and gave it a disagreeable odour and taste. When the bloom became dense, the dam reservoir was treated with I ppm of copper sulphate (Bourke et al., 1983). Clinical injury among consumers on that water supply was reported the following week. In subsequent investigations, the organisms from the dam were cultured and administered to mice. Mice treated with Raphidiopsis (Cylindrospermopsis) raciborskii culture slowly developed (over several days) widespread tissue injury involving the gastrointestinal tract, the kidney and the liver (Hawkins et al., 1985). The widespread tissue damage and delayed effects are quite different to those following Microcystis aeruginosa administration (Falconer et al., 1981). Subsequent monitoring of the blooms in the dam - well after the outbreak - identified R. raciborskii as the cause of the blooms, with seasonal cell concentrations of up to 300 000 cells/mL of water. This organism did not form scums and has the highest cell concentrations well below the water surface. In order to reduce bloom formation, the responsible authorities later introduced destratification of the reservoir (Hawkins & Griffiths, 1993). Subsequent research on toxins produced by R. raciborskii has identified the cytotoxic alkaloid cylindrospermopsin.
