Water companies and the DWI were angered by claims made at a recent meeting of the Society for Applied Bacteriology (SAB) that they were ignoring the potential health risks posed by algal toxins in water supplies (ENDS Report 236, p 11 ). In fact the industry, the DWI and the National Rivers Authority have collaborated in sponsoring a four-year research programme through the Foundation for Water Research (FWR).
The programme has addressed many of the key issues surrounding the toxins, including the development of analytical methods to detect them and an assessment of their susceptibility to water treatment processes. Toxin levels in water supplies have also been monitored during algal blooms.
The latest report sponsored by the DWI, and shortly to be published by WRc, is a survey of toxin levels in six water supplies during August and September this year. Blue-green algae were present in four of the sources at levels up to 18 million cells per litre. But analyses failed to find the toxins microcystin-LR or anatoxin-a in treated water above a detection limit of 1µg/l.
An earlier survey in 1993 looked at toxin levels in ten water supplies from across the UK. All contained high levels of blue-green algae, and measurable levels of anatoxin-a were found in three supplies from Northern Ireland which contained blooms of the alga Oscillatoria. Analyses found up to 0.7µg/l of anatoxin-a in raw water and up to 1.3µg/l in treated water. Preliminary toxicity data suggested that these levels are not of health concern, and this finding has been confirmed by a more complete report on the toxicity of anatoxin-a to be published shortly by FWR.
In 1992, a survey of seven water supplies found microcystin-LR levels of 0.2µg/l in raw water from a Scottish loch and 8,500µg/l in a sample from a reservoir in northern England. The latter was a surface water sample containing a dense bloom of the alga Microcystis. Attempts to centrifuge the alga out of solution may have caused cell lysis, the survey concluded, leading to very high toxin levels.
With the exception of the last figure, none of the surveys have found levels as high as the 10-16µg/l detected in supplies from Rutland Water, Leicestershire, in 1989. This may be a result of less favourable weather for blooms in recent years, or the relatively crude analytical methods available in 1989 which may have produced a high result.
For microcystin analyses, a WRc method which detects only microcystin-LR rather than all microcystins was used. One of the points raised at the SAB meeting was that this method could underestimate microcystin levels. There is some support for this in the 1992 survey, which noted that other microcystins appeared to be present in the sample containing 8,500µg/l of microcystin-LR. The survey concluded that the total microcystin level in this sample could be as high as 10,000µg/l.
The FWR research programme has also included a toxicological assessment of microcystin-LR. Mice were found to be more sensitive than rats, but displayed a no observed effect level for tissue damage of about 40µg/kg body weight. In foetal toxicity surveys, the study found no effects on foetuses even when the mother showed symptoms of toxicity, although this work needs to be repeated on other animals.
In terms of a risk assessment, a dose of 40µg/kg of microcystin-LR would be 250 times the dose received by an infant consuming 0.75 litres/day of water containing 1µg/l of the toxin, the study concluded. The potential tumour-promoting action of microcystin-LR has yet to be assessed.
In a letter to ENDS, the DWI said that the studies had shown that toxins are "an important consideration for recreational use of water. However, there is no evidence that algal toxins present any risk to health via water supplies".
The DWI also defended the analytical methods for toxins recommended by the Standing Committee of Analysts which were criticised at the SAB meeting. The methods "receive a far higher level of scrutiny than the so-called peer review process", the DWI said, adding that their validity had been confirmed by five separate laboratories in over 600 tests.
Advanced water treatment processes are capable of removing both microcystin-LR and anatoxin-a from drinking water, the FWR research programme has found. Filtration with granular activated carbon was very effective at a contact time of 15 minutes. This is a moderately high application rate similar to that needed for pesticide removal but higher than that used for routine taste and odour improvement.
Dosing clarified water with oxidising agents was also effective in removing or degrading toxins. Potassium permanganate applied at 1mg/l was effective in degrading 90% of toxins, and ozonation destroyed 90% of toxins at a dose of 2mg/l. However, chlorine disinfection was ineffective except at low pHs, and UV light achieved only about 30% degradation at the doses typically.