Food Surveillance Information Sheet Number 144

ANNEX C - Comments from the UK Tea and Coffee Trade

The UK Tea Trade Technical Committee (UKTTTC), the British Soluble Coffee Manufacturers' Association (BSCMA) and the Roast and Ground Coffee Association (RGCA) have commented on the results of this survey, as reported in the original version of this Information Sheet. Their comments are summarised and discussed below.

Factors affecting caffeine concentrations in tea

The UKTTTC has noted that caffeine levels in tea as consumed are affected by both the broad range of preparation methods open to consumers and the variation in the caffeine content of the natural raw material. For example, the weight of dry tea, volume of water, temperature of the water, brewing time, extent of stirring and, with bagged tea, the extent of squeezing of the bag, can all affect the caffeine content of the resultant drink. As well as variation introduced by consumer preparation methods, there is also an inherent variation in tea itself. Tea on sale to the consumer is generally a blend of teas from a range of countries. Although standardised for taste, the particular blend of teas is likely to vary according to economic factors and the growing conditions affecting each individual harvest. Caffeine levels may therefore vary slightly between batches.

Tea bags

The UKTTTC has queried the method which was used to prepare tea bag infusions (1 standard tea bag, infused for 5 minutes, without agitation in 200 ml water) on the grounds that this does not reflect typical consumer practice. These conditions had been selected since they were in line with those used in an earlier UK survey of caffeine.⁹ Consumer surveys commissioned by UK tea manufacturers have shown that more than 90 per cent of domestic consumers use tea bags. An estimated 80 per cent prepare tea directly in the cup or mug, usually using a "stir and squeeze" technique and 200-300 ml water. Such infusions are typically carried out for less than 1 minute (mean, 52 seconds) with only 2-3 per cent of consumers leaving the tea bag to infuse for 3 minutes or more. For tea made from tea bags in a teapot, the duration of the infusion is generally longer, with an average of approximately 2 minutes and about 15 per cent of consumers brewing for 3 minutes or more.

Information provided by the UKTTTC suggests that caffeine concentrations reported in this survey would have been around 30 per cent lower if the infusions had been terminated after 1 minute. A marginally increased water:tea ratio would also have lowered the resultant caffeine concentrations. The UKTTTC estimates that a typical tea bag infused in 230 ml of water for 1 minute with the use of a stir and squeeze technique will give a drink with a caffeine concentration of around 200 mg/l. The UKTTTC has also reported that brewing times between 40 seconds and 6 minutes result in a range of caffeine concentrations of 162-278 mg/l. The conditions used in this survey, therefore resulted in caffeine concentrations at the upper end of the range which may be experienced by consumers.

Loose teas

The UKTTTC has expressed concern at the low caffeine extraction achieved with the three loose tea products in the survey. Tea manufacturers report that the total caffeine content of loose teas is approximately 3 per cent and that the majority of this would be expected to be extracted into the tea infusion under the conditions used in the survey.

Unilever Research, on behalf of the UKTTTC, has recently analysed the three loose tea brands which were sampled in this survey, comparing the infusion conditions used in this work (1.6 g infused for 5 minutes in 200 ml water, without stirring) with those considered by the tea trade to be representative of typical consumer practice (3.2 g tea infused for 5 minutes in 500 ml water). The results are shown below in Table C1.

Table C1: Concentrations of caffeine in loose tea (supplied by the UKTTTC).

^aAverage extractable caffeine content (mg/g tea) in brackets

As part of this work, samples were analysed to test the effect of stirring on the extraction of caffeine from loose tea. This showed that constant stirring results in caffeine concentrations which are approximately 25 per cent higher than in unstirred infusions (Table C1).

There is no obvious explanation for the difference between UKTTTC's results for loose tea infusions and those obtained in this survey. The latter are consistent with some of the previous reports in the scientific literature,^4,11 but the results of the UKTTTC suggest concentrations approximately 50 per cent higher (Table C1).

The caffeine concentrations reported by the UKTTTC using conditions considered to be more typical of consumer practice (3.2 g infused for 5 minutes in 500 ml water) were between 129 and 168 mg/l (Table C1). The UKTTTC has also reported a range of caffeine concentrations of 63-364 mg/l for loose tea infusions prepared using a variety of methods.

Instant teas

The original Information Sheet reported the levels of methylxanthines resulting from the dissolution of 1.6 g of instant tea in 200 ml water. The UKTTTC pointed out that a typical consumer would not use this ratio of product to water and that the labels on the packets would suggest portions which are lower in three cases (PG Tips instant tea; Sainsbury's Red Label; and Tetley Instant Tea Granules) and higher for two others (Lift instant lemon tea and Typhoo QT). The results for these products have been adjusted accordingly (Tables 2a and A4).

Filter and percolated coffee

Both the BSCMA and the RGCA have commented that the weight of coffee grounds used to prepare filter and percolated coffee (2.6 g in 200 ml of water) was rather low. This sample weight had been selected because the labelling on the majority of the products specified the use of a dessertspoonful of coffee per cup. On investigation, it was found that one level dessertspoonful (10 ml) of ground coffee weighed approximately 2.6g.

The BSCMA has supplied information which indicates that consumers preparing coffee in an electric filter machine typically use between 3 and 15 g of ground coffee per cup, with a mean of 7.6 g. The conditions used in this survey are therefore at the bottom end of this range. Assuming a linear relationship between the weight of coffee and the amount of caffeine extracted, the range of caffeine concentrations which would be expected using 7.6 g in 200 ml can be calculated, from these survey results, to be 306-628 mg/l. Assuming a weight of 15 g in 200 ml, the corresponding caffeine concentrations would be 606-1240 mg/l. These calculations confirm the wide potential variation in the caffeine content of filter and percolated coffee as consumed.

Conclusions

The information provided by the UK tea and coffee trade associations confirms that consumers prepare tea and coffee under a variety of conditions and, as a result, the range of caffeine concentrations in the products as consumed will be very wide. As in other surveys, the results reported in the Information Sheet have been derived using standard conditions which will overestimate the amount of caffeine consumed by some individuals and underestimate that by others.

In the case of infusions prepared using tea bags, the concentrations reported in the survey are probably towards the upper end of the range, since most consumers will use shorter infusion times and/or a larger volume of water. Similarly the results for filter and percolated coffee are likely to be at the lower end of the range experienced by coffee drinkers, since most consumers are reported to use a larger amount of coffee grounds. Therefore, although the survey indicates that caffeine concentrations are higher for tea than for filter or percolated coffee, this may not be the case for drinks made under "average" conditions.

The results reported for loose teas, while in line with some other literature reports, are lower than those obtained elsewhere and by the tea trade. Although many factors are known to affect the extraction of caffeine from tea leaves, there is no obvious explanation for this difference.

Return to Index to Surveillance Information Sheets, 1998 page
Go to top of page

These pages were last updated on 27th February 1998