Food Standards Agency UK - Survey of Colours in Sauces (Number 03/00)

• 0.9 per cent of samples (2/227) contained levels of individual colours that exceeded the maximum permitted limit of 500 mg/kg (milligrams per kilogram) in the sauce.
• The Food Standards Agency and Local Authorities have taken action to ensure that consumers are not exposed to high levels of colours in sauces.

There have been reports in the press that levels of red and yellow colours in sauces used on some foods are above the permitted limit.^1-7 These reports followed work by Local Authorities. It was not clear whether this was primarily a local issue or was more general.

The survey described here was conducted to determine the extent to which levels of colours in sauces used on foods in Great Britain exceed the maximum permitted limit. The use of colours in foodstuffs is controlled in Great Britain by The Colours in Food Regulations 1995 (parallel legislation exists in Northern Ireland).⁸ In these regulations and hence in reporting this survey 'colour' means a food additive that is used or intended to be used for the primary purpose of adding or restoring colouring in a food. The following colours were included in this survey: E102 (tartrazine), E110 (sunset yellow FCF, orange yellow S), E120 (carminic acid, cochineal, carmines), E122 (carmoisine, azorubine), E124 (ponceau 4R, cochineal red A), and E129 (allura red AC). These may be used singly or in combination in sauces, seasonings (e.g. curry powder, tandoori), pickles, relishes, chutney and piccalilli in each case up to the maximum level specified for each food of 500 mg/kg. During the survey it was apparent that clarification is needed on the control of combinations of colours. Different limits may apply for colours when they are used in combination. The use of materials such as saffron, turmeric, and paprika falls outside the scope of these regulations.

Sampling
Three hundred and twenty duplicate samples were purchased by Ventress Technical Services (VTS) Ltd. in 1998 and 1999, from ten localities in England, Scotland and Wales. In each locality 32 paired samples were purchased (8 from supermarkets, 12 from butchers and 12 from takeaways). Samples that were in poor condition or otherwise unsuitable were replaced by samples purchased in the Cambridge area.

Samples were selected that were distinctly coloured red, yellow or orange. All samples were uniquely bar-coded at the time of purchase and their details were entered on individual sample log sheets. One set of samples was sent for analysis by Reading Scientific Services Ltd. (RSSL) and one retained by VTS Ltd. should others wish to undertake their own analysis of the samples.

Analysis
Samples that contained a distinct amount of sauce were extracted using method 1 (quantitative): a 2.5g aliquot of well-mixed sauce was diluted to 25 ml with 4 per cent methanolic ammonia solution, shaken vigorously and left to stand for 30 minutes. Samples that did not contain sufficient free sauce were extracted using method 2 (comparative): a weighed portion of the sample was initially rinsed with water followed by methanolic ammonia solution and the washings were combined and made up to 100 ml. The filtered supernatants from each extraction (method 1 or 2) were evaporated to dryness and re-dissolved in 1 ml 0.02M ammonium acetate solution. These extracts were filtered through a glass fibre membrane prior to analysis.

The extracts were analysed in 29 analytical batches of 10-15 samples each, using high performance liquid chromatography (HPLC) methodology with diode-array detection. The colours of interest were identified by retention time and their identity was confirmed by diode-array spectral matching. The concentration of each colour in the samples was calculated as a proportion of the peak area produced by the corresponding external standard (corrected for colour content). The limit of quantification (LOQ) for each colour was 1 mg/kg, except for E120 which had a LOQ of 10 mg/kg.

Quality Assurance
Standard solutions of individual colours were combined to produce a mixed calibration standard with a concentration of approximately 50 mg/l for each colour. This concentration was equal to that present after the preparation of a sample that contained a single colour at the permitted limit of 500 mg/kg. Calibration standards were run at the start and end of smaller batches. An additional standard was included in the middle of batches that contained more than 15 samples. At least one aliquot of reference curry sauce with added, known amounts of colours was analysed with each batch of samples to obtain recovery data. The reference curry sauce was previously tested to confirm the absence of the colours of interest and that it was a matrix comparable to the samples under surveillance. When the analytical result for a sample was in excess of 500 mg/kg, a further sample extract was prepared and analysed to obtain a confirmatory result. The maximum difference in analytical results between initial and further extracts of samples was 14 per cent.

Reporting
Quantitative results are reported for 227 samples in Table 1 as mg colour/kg sauce. In the other 93 samples it was not possible to analyse the sauce as it had become almost totally dried onto or absorbed into the meat. An estimate of colour levels was made for these samples based on the following assumptions:

• using extraction method 2 the entire content of colour was washed from the food; and
• any colour(s) had originally been applied in 5 ml of sauce per 100 g of product. The levels estimated for samples with dried on or absorbed sauce (Table 2) are included in this report for completeness.

Recoveries of added colours from the reference curry sauce were close to 100 per cent, except for E120 which gave consistently low recoveries (mean: 45 per cent). As recoveries were close to 100 per cent, except for E120 which was not found in any sample, the results have not been corrected for recovery. Despite the low recovery of E120 from the reference curry sauce, it would have been found at levels greater than 10 mg/kg in the sauces analysed had it been present. Where no colours were found at or above the limits of quantification, this is reported as NQ (not quantifiable). The combined limit of quantification (15 mg/kg) was well below the respective maximum permitted limit for these substances, of 500 mg/kg. Thus a limit of detection, which would have been less than the limit of quantification, was not applied in this survey.

Brand names were reported as this survey was carried out in accordance with guidelines for reporting survey results published in the Food Safety Information Bulletin in September 1997. The absence of a particular brand from Table 1 and Table 2 means only that the brand was not included in the survey.

Quantitative results
Two samples out of 227 had unexpectedly high levels of individual colours. The samples were of a chicken tikka masala (containing 2,054 mg of E110 per kg) sampled in central London, and a nishan tandoori paste (containing 1,688 mg of E124 per kg), sampled in Birmingham (Table 1). Both of these results were well above the maximum permitted limit. If the paste were to be diluted with yoghurt as suggested by the manufacturer's 'Chef's tip' on the label the level would be slightly below 500 mg/kg. However, this sample was still considered to be in excess of 500 mg/kg as the 'Chef's tip' cannot be regarded as instructions for use since it is a suggestion as to how the customer may like to use the product. Re-extraction and re-analysis of these two samples gave similar results (2,265 mg E110/kg in the chicken tikka masala sampled in central London, and 1,661 mg E124/kg in the nishan tandoori paste sampled in Birmingham).

Three other samples contained total levels of tested colours that were close to the permitted limit: 581 mg/kg in a sample of pork chops obtained in Manchester and 587 and 564 mg/kg in samples of chicken tikka from Southampton and Cardiff respectively (Table 1). Re-extraction and re-analysis of these samples gave respective values of 567, 603 and 542 mg/kg. The sample of chicken tikka from Southampton contained a mean level of 586 mg E110 per kg. The other sample of chicken tikka and that of pork chops did not contain more than 500 mg/kg of a single colour. Since there is some doubt about what limit to apply to mixtures of colours, and there was some variation between the results for extracted and re-extracted samples, it was not possible to state with absolute certainty that the maximum permitted limit had been exceeded for these three samples.

There was no geographical pattern in the amounts of colours found. To investigate any significant underlying trends the data were subjected to statistical analysis using chi-square calculations for multinomial probabilities. There was not a significant variation in the presence of the colours (p greater than 0.95).

Estimated levels for samples with dried on or absorbed sauce
One sample out of 93 might have had a level of one colour that exceeded 500 mg/kg (estimated at 711 mg E110 per kg in a kebab sampled in Manchester; Table 2). However, as it was not possible to ascertain how much sauce was originally present in the sample no conclusions can be drawn.

The results show that in general colours were being used in sauces in accordance with the legislation. Indeed no health effects would be expected in an individual's lifetime from the levels found in this survey. However there was evidence that the law had been broken in the case of the two samples with unexpectedly high levels of individual colours. The results for these samples were sent to the respective companies and to the respective local enforcement authorities.

The survey did not find differences in the amounts of colours in sauces from one part of the country to another. Local Authorities have found some variations within their areas that they have followed up to ensure that colour levels are within the permitted limit. Following a local survey in Kent the provision of information and advice about the use of colours generally led to lower levels. This action is being taken in other parts of the country where samples containing more than the permitted limits of colours have been found.

The law on colours in sauces needs to be made clearer in two respects:

• The addition of colours to most meats is not permitted, whilst the use of colours in sauces for colouring meats is permitted. Guidance is required for enforcement authorities on identifying retail products coloured in these different ways.
• The control of colours used in combination needs to be clarified. Part II of Schedule 5 of The Colours in Food Regulations 1995 states that specified colours (including those in this survey) 'may be used singly or in combination in the foods listed in Part III of this Schedule,' (which includes sauces) 'in each case up to the maximum level specified for such food'. .⁸ It is not clear what the word 'each' refers to - colours or foods.

As the law on colours in food is harmonised in the European Union these points have been taken up by the Food Standards Agency with the European Commission.

1. Why a red hot sauce can leave you off colour. The Daily Mail 14 July 1997.
2. Red alert in the Rhondda. Trading Standards Review News vol. 105; September 1997.
3. Health scare over curry colour. The Times 9 July 1998.
4. Red peril of the chefs who are dyeing to serve you a curry. The Daily Mail 9 July 1998.
5. Curry colourings threaten health. The Independent 11 December 1998.
6. Food colouring in curries - Watchdog Health Check. BBC Online. Accessed 25 January 1999.
7. Off colour. Trading Standards Review News vol. 107; October 1999.
8. The Colours in Food Regulations 1995. (S.I. 1995 No. 3124).

Dr David Watson
Food Standards Agency
Room 516c Aviation House
125 Kingsway
London WC2B 6NH

Tel: +44 (0) 7276 8537
Fax: +44 (0) 7276 8514
E-mail:david.watson@foodstandards.gsi.gov.uk

A copy of the full report of this survey has been placed in the library in Aviation House, 125 Kingsway, London, WC2B 6NH (Tel. No. +44 (0)20 7276 8181). If you wish to consult a copy please contact the library for an appointment giving at least 24 hours notice or alternatively copies can be obtained from the library; a charge will be made to cover photocopying and postage.