Water water everywhere…

Contamination events can happen at any time – day or night – and analytical laboratories must act fast to produce rapid, accurate contaminant identification, how can they be sure their result is the right one?

Accidental or malicious contamination of potable water supplies – although rare events – could have severe implications for both the general public and the water industry. These include potential adverse impacts on public health and financial impacts on the company through necessary remedial action, possible significant fines following prosecution and legal actions from affected consumers, all of which could result in irreparable damage to a company’s reputation.

For example, the 1988 Camelford water pollution incident – which involved the accidental contamination of the town’s drinking water supply with aluminium sulphate – is perhaps the UK’s worst incident of its kind to date. A range of short-term health effects were found in many people who came into contact with the contaminated water. The long-term implications for those who were affected remain unclear.

Obviously, the consequences arising from laboratories not identifying or misidentifying a potential water contaminant could be very serious. Add to this the sheer scale of the analytical challenge involved – the Chemical Abstracts Service (CAS) Registry currently contains more than 54 million organic and inorganic substances, and approximately 12,000 new substances are added every day – and we begin to see the importance of producing rapid, accurate analytical results. Regular assessment of the competence of laboratories to rapidly identify such contaminants is therefore vital.

Independent proof of competence of analytical laboratories is demanded on an increasingly frequent basis from both regulatory bodies and customers. Laboratories need to know how well they are performing against objective standards and how well their performance compares against that of others. Participation in proficiency testing (PT) schemes enables laboratories to obtain an objective and confidential method of assessing analytical accuracy and competence. Proficiency testing is also a quality assurance procedure commonly used by accreditation bodies to assess the ability of laboratories to competently perform tests or calibrations for which accreditation is held. However, it is not a substitute for the external audit process that is accreditation.

PT schemes operate by sending out test materials to several laboratories for evaluation. It is not unusual for 50-250 laboratories to participate in any one scheme on a regular basis. Participants receive a sample spiked with a known level of the substance to be measured, and are free to use whatever method they wish to analyse for that substance. The collated results are used to derive the best estimate for the “true” level of the substance (also known as the assigned value) and then each result is assessed against a predetermined allowable variation (based on best practice) around this “true” level. The difference between each result and the assigned value is compared to the target standard deviation. The end product of the performance assessment is a standardised statistic known as a z-score.

The Laboratory Environmental Analysis Proficiency (LEAP) scheme, initially set up in 1995 to serve the water environmental sector, is run by the Food and Environment Research Agency (Fera). Scientists at Fera also organise a scheme based on LEAP, which tests the ability of laboratories to identify an unknown substance(s) in a potable water sample during a simulated emergency contamination situation. The LEAP Emergency PT scheme is aimed at providing participating laboratories with information on their ability to produce fit for purpose, timely results during an emergency potable water contamination incident. With typical PT schemes, the participating laboratories know what they are being asked to measure. However, emergency schemes are based around the rapid assessment of an unknown contaminant in a sample, under an emergency contamination scenario. Laboratory staff are totally unaware of when the sample will arrive in order to try and simulate a real emergency incident

The LEAP scheme organises two “exercises” per year, where potable water is spiked with up to 6-7 inorganic and organic contaminants. Approximately 250  contaminants used for spiking since 1995 have previously included phenol, dimethoate, 1,4-dioxane, phenyl mercuric acetate, thallium, methanol and imidacloprid. For some contaminants the laboratory will also detect the contaminant breakdown products. In 2007 a radioactivity sample for gross alpha and beta analysis was also introduced. A blank uncontaminated water sample is supplied. These samples are sent to the participating laboratories, without prior warning, together with a contamination scenario. A typical example is given below:

Participating laboratories are asked to return results as soon as they become available. Questions to be answered include: •    Is there any significant contamination? •    If so, what is the approximate concentration(s) of the various contaminants? •    What can the water be used for? •    What analytical methods were used? •    Were any screening tests used?

Each laboratory should judge its performance by how quickly it identified the various contaminants; how accurately it estimated the concentration of each identified contaminant; how many contaminants present were not identified and were any contaminants that were not in the sample falsely reported as present

Once the exercise has closed, laboratories are informed of the contaminants present and their concentrations. A full confidential report on the participant’s performance is provided in due course. The reports contains details of what contaminants the laboratory identified and their concentration, the time each contaminant result was reported to Fera, the analytical and screening methods used, what the water can be used for and any other contaminants reported (false positives). Discussion forums are held at Fera under Chatham House rules after the closing date for each exercise to give participants a chance to discuss the outcomes of the exercise and share information and views on the best routes to detect contaminants.

From the exercises run to date Fera’s scientists have found that: •    Inorganic chemicals and metals are usually correctly identified •    Non-polar organic substances have poorer detection rates •    Polar organic substances such as methanol are often not detected •    Contaminant concentrations are often significantly under or over estimated

Participation enables the laboratory to demonstrate its ability to identify unknown contaminants by the most efficient and timely means; provides useful quality information e.g. testing new screening techniques and to show where further development work is needed; allows the laboratory to measure its performance against others; and serves as an ongoing test of laboratories’ emergency testing capabilities.

Professor Clive Thompson, Chief Scientist at ALcontrol Laboratories stated “The last 15 years has clearly demonstrated that this scheme has resulted in a significant improvement in the performance of the participating laboratories. This has been found to be an ideal cost-effective scheme format for proficiency testing of very low frequency and very high impact water contamination emergency events.”

Fera is keen to ensure that this novel scheme is fit for purpose and we have found that with the water emergency test, the participants and other interested bodies (such as the Drinking Water Inspectorate) have seen this type of exercise as highly educational. Taking part in these tests has led to participants making a marked improvement in detecting unknowns in water.

Other schemes like this are being developed for the analysis of contaminants in foodstuffs under simulated incident conditions. Like the potable water industry, food contamination is occasionally deliberate, but can also arise through taints in food or the formation of chemicals through processing. A number of high profile issues have highlighted the need for broader determination of the composition of the foodstuffs, most recently the fatalities that arose from the contamination of milk with melamine. Other examples that were not anticipated include the formation of acrylamide in cooked and heat-processed foods and the presence of benzene in mineral water.

Emergency facilities for the detection of chemicals in food and water are offered by a range of contract laboratories. In order to establish the efficacy of new technologies in comparison to existing approaches a new emergency proficiency testing scheme for the food industry will be launched soon by the Fera proficiency testing team, providing analytical laboratories and their customers with robust assurance in the results produced.

The Author Dr Alison Wilson is the Media Officer at the Food and Environment Research Agency

Contact: t:01904 462380 e:alison.wilson@fera.gsi.gov.uk