A Technical Guide
for Wine Producers


Ethyl Carbamate in South African Wine

Michael Waldner

Michael Waldner1,2 and Ockert Augustyn1,3
1. ARC Infruitec-Nietvoorbij, Stellenbosch
2. Current address: Woolworths Foods Laboratory, Cape Town
3. Corresponding author

Key words: Wine, Ethyl Carbamate, Urea, South Africa



Ethyl carbamate (EC) occurs naturally in all fermented foods and beverages. Because EC is a proven carcinogen in certain animals the discovery of its natural occurrence in products for human consumption caused quite a stir. Initial surveys indicated varying concentration levels in different products with some alcoholic beverages (particularly stone fruit distillates) containing extremely high levels. Internationally, producers of alcoholic beverages have been striving to lower EC levels as far as possible. This activity was spurred on by Canadian legislation (1986) which set maximum permissible limits of 30 parts per billion (ppb) for natural wine, 100 ppb for fortified wine, 150 ppb for distilled spirits and 400 ppb for fruit brandies to be sold within their borders. In response to the Canadian actions, the US Food and Drug Administration (FDA) established a voluntary target of 15 ppb for natural wine and less than 60 ppb for fortified wine produced in the USA. Additionally, the FDA notified all countries exporting wines to the United States that they must develop programmes to meet these targeted levels.

The purpose of this Winetech supported project was therefore to determine the EC status of local wines and factors affecting these concentrations.

Origin of wine ethyl carbamate and factors affecting concentration

In general, red wines develop higher concentrations of EC than white wines. The bulk of EC present in wine is formed by the spontaneous reaction between urea and ethanol. Ethyl carbamate formation increases over time with the reaction rate being exponentially accelerated at elevated temperatures. Urea is formed when the wine yeast metabolises arginine, a major alpha-amino acid in grape juice available to yeast. This reaction is yeast strain dependent. Yeasts differ in their ability to produce urea and to re-use urea secreted into the must / wine. The ability to utilise secreted urea is obviously also affected by the general nitrogen status of the must. Lactic acid bacteria also metabolise arginine and liberate citrulline, an amino acid, which then reacts with ethanol to form EC. From the above it is clear that high arginine musts fermented by high urea producing yeasts will invariably contain high levels of urea, resulting in potentially elevated EC concentrations. Over-fertilised vineyards, in general, will yield wines with higher urea and EC concentrations.

Samples analysed

Ethyl carbamate was determined (coupled gas chromatography / mass spectrometry) in more than 4300 different wines. These wines came from the 1997 and 1998 Young Wine Shows, the 1999, 2000 and 2001 Veritas Shows, individual producers or were purchased from retail outlets as needed. Upon arrival at Nietvoorbij all wines were stored below 18C until analysis. Urea was determined in a number of wines from the 1997 Young Wines Show. Total alpha-amino nitrogen content of a number of wine grape cultivars was determined on material gathered in a rootstock trial. Ten different rootstocks were used in the trial, each being grafted to eight different scions (cultivars). Although it was attempted to harvest all samples at approximately the same stage of maturity (B), this was not always possible.

Results and discussion

Young Wine Show Wines

The urea concentrations of 312 wines from the 1997 Young Wine Show are summarised in Table 1. Data for 308 red wines from the 1998 Young Wine Show showed very similar trends, and are therefore not shown. Kodama et al. (1994) clearly indicated that EC formation is closely related to urea content. These authors also determined that urea levels should be kept below

2 mg/L if EC levels in wine are to stay below the USA voluntary limit of 15 ppb. For wines to safely remain below the Canadian limit they indicated urea concentrations of < 5 mg/L. Both deductions are only valid if wines are stored at or below 20C. The data in Table 1 show that generally red wines have higher concentrations of urea than white wines. Grape cultivar also seems to affect urea concentration levels.

Wines from the 1997 and 1998 Young Wine Shows were also analysed for EC. Only data for red wines are recorded in table 2. White wines (1997) contained much less EC than the red wines at this early stage e.g. Chenin blanc: mean 5.02 g/L (n = 10); Chardonnay: mean 4.77 g/L (n = 33) and Sauvignon blanc: mean 3.47 g/L (n = 12). Mean EC concentrations in the 1998 red wines are appreciably lower than those found during 1997. One of the reasons contributing to these results could be that analysis of the 1997 wines started six month after the show whereas the 1998 analyses were completed within two months after the show.

Tables 1 and 2 show that the urea concentration varies from 1,24 to 5,23 mg/litre and the EC concentrations from 2,82 to 8,24 ppb. Table 3 indicates that Pinotage juice consistently contains the highest concentrations of alpha-amino nitrogen of the red cultivars studied. However, why white wines tend to have such relatively low EC concentrations when considering the high alpha-amino nitrogen levels recorded during 1999, is not clear at present. Possibly the differences in the production process play a major role in this regard.

Data recorded in Table 4 add additional food for thought. Certain rootstocks clearly tend to induce higher concentrations of alpha-amino nitrogen. In cultivars that naturally have high levels of available alpha-amino nitrogen, over-fertilisation should be avoided at all cost. In such circumstances rootstock choice could also exaggerate or diminish potential problems.

Veritas Wines

Perusal of the data presented in Table 5 gives a clear picture of the EC status of unfortified South African wines. Clearly local white wines present very few problems in this regard. The bulk of the 1126 white wines represented in this table had EC levels <10 ppb. Chardonnay wines from the 2000 show are a conspicuous deviation from this norm with only 83% having EC concentrations below 10 ppb. However, all values for all wine types were higher in this data set for reasons not immediately evident.

Although, as expected, wines from red cultivars tend to have higher EC concentrations than those generally found in white wines [table 5], this is no cause for undue concern. If between 34% and 89% of red wines [depending on cultivar and vintage - table 5] can have EC concentrations below 10 ppb, there is no reason why levels in the rest cannot be lowered appreciably by application of current knowledge on EC and its precursors.

Wines entered in the Veritas Shows are typically between one and four years old. As EC concentration increases with wine age, the percentage of "older" wines in a batch could appreciably affect the mean concentration for the group.

For some of the show wines that exhibited really high EC concentrations we purchased reference samples from the producers. In all cases these reference wines recorded appreciably lower EC concentrations than the show wines. This does seem to suggest that wines sent to the Veritas Show were subjected to elevated temperatures somewhere along the line. A similar scenario was found for a 1999 Merlot rejected by Canada late in 2001. The Canadians reported a surprising 39 ppb EC. Our own analysis of stock from the same batch supplied by the producer only indicated 24 ppb. We are confident that the difference is the result of the wine being subjected to elevated temperatures during its journey to Canada. This confidence is supported by our examination of a second wine also rejected by the Canadians during this period. Here our analysis of local reserves of the 1999 Pinotage confirmed the 32 ppb found by the Canadians exactly.


That we need to address the matter of making wines with EC concentrations as low as technically possible, is emphasised by the following: Although the Canadian legal limit for natural wine remains at the 1986 level, they currently do not accept wines within 80% of that limit. This means that wines having EC concentrations of 24 ppb are not going to be acceptable. If the FDA applies the same principle (which is based on the fact that EC formation is an ongoing process) their limit of acceptability will drop to 12 ppb.

The modern consumer is quite rightly much more health conscious than his / her forefathers. In line with these concerns, legislation aimed at protecting consumer health is constantly being updated and made more stringent. This is particularly true for our sophisticated first world export markets. Reducing EC concentrations in our products is therefore a prerequisite if we want to continue to export to these markets in the long term. Concern for consumers at home should also not be less than that reserved for those in our export markets and the general lowering of levels of potentially dangerous substances in all our products should remain a common goal.

The real health risks posed by EC in local (and other) wines are very difficult to assess at this time. As mentioned earlier, EC is a proven carcinogen in certain animals. This is true when it is administered in high doses (mg/kg body weight). In addition, preliminary evidence indicates that the carcinogenic action of EC in mice is inhibited if it is injested in wine or alcohol solution (Stoewsand et al., 1991). Therefore, while our efforts to lower EC levels in all our products should continue unabated, the current status does not seem to be reason for significant health concerns.

How does one limit ethyl carbamate concentrations in wine?

From the above and the manual by Butzke & Bisson (2002) the following:

  • Do not over-fertilise. Not only does it waste money, but it elevates assimilable nitrogen levels in musts thus ultimately leading to elevated residual urea levels in the resultant wine. Also consider cover crops in case of potentially problematic cultivars. Do not use crops that add nitrogen to the soil. Remember the data in Table 4 - rootstock choice may be important. Do not add unnecessary amounts of diammonium phosphate to musts as a matter of course (know the N-status of your musts).
  • Choose the correct yeast. As pointed out earlier yeasts differ in ability to re-use urea formed by their utilisation of arginine. Preliminary data in Table 6 illustrate this fact for five yeasts currently in use locally [Data from a new study examining urea production by 12 local yeasts in three cultivars over two vintages will be published later]. Correct choice of yeast for the must at hand will significantly lower EC levels.
  • Spontaneous malolactic fermentation may contribute to elevated EC levels. Use commercial strains that do not produce high levels of citrulline.
  • Lees contact. Although Butzke & Bisson (2002) report that this method does not dramatically impact on EC levels, they also mention that this fact has not been thoroughly tested. Preliminary results from our laboratory would seem to suggest that ageing on lees can indeed contribute to EC levels. Possibly a technique to be avoided in the case of high alpha-amino nitrogen containing cultivars?
  • Storage and transport. Ethyl carbamate formation is exponentially increased at elevated temperatures. Maintenance of the correct cold chain [from production through shipment to retail venue] is crucial.
  • Enzymatic removal of residual urea. Urease (commercial preparations are available) can be used to reduce urea. However, general conditions in wine may limit effectiveness. Use of this enzyme must be evaluated in each wine to confirm that it is active.

Literature cited

Butzke, C.E. & Bisson, L.F., 2002. Ethyl carbamate preventative action manual. http://vm.cfsan.fda.gov/~frf/ecaction.html

Kodama, S., Suzuki, T., Fujinawa, S., Teja, P. & Yotsuzuka, F., 1994. Urea contribution to ethyl carbamate formation in commercial wines during storage. Am. J. Enol. Vitic. 45, 17-24.

Stoewsand, G.S., Anderson, J.L. & Munson, L., 1991. Inhibition by wine of tumarigenesis induced by ethyl carbamate (urethane) in mice. Food Chem. Toxicol., 29, 291-295.

For further information contact: O.P.H. Augustyn at tel (021) 809-3010, fax (021) 809-3400 or e-mail augustyno@arc.agric.za.


Ethyl Carbamate (EC), a proven carcinogen in certain animals, occurs naturally in all fermented foods and beverages. The bulk of EC present in wine is produced from the spontaneous reaction between urea and ethanol. Urea is formed when wine yeasts metabolise arginine, a major alpha-amino acid present in grape juice. Levels of urea are yeast strain dependent and also affected by the nitrogen states of the must. More than 4300 South African wines were analysed to get an overview of the occurrence of EC. In general it can be stated that EC levels in red wines tend to be higher than in white wines. Grape cultivar, scion / rootstock combination and wine storage temperature all affect the level of EC found in a particular wine. Advice is given on how to minimise EC levels in local wine.

Table 1: Urea concentration (mg/L) in young wines obtained from the 1997 Young Wine Show
Table 2: Ethyl carbamate concentrations (ppb*) in red wines obtained from the 1997 and 1998 Young Wine Shows
Table 3: Total alpha-amino nitrogen content (mg/L) of grape cultivar juice during 1999, 2000 and 2001
* Mean value for each cultivar on 10 different rootstocks
Table 4: The total alpha-amino nitrogen content of grapes obtained from different scion / rootstock combinations in 1999
* Mean value for same rootstock attached to eight scion cultivars
** Means with the same letter are not significantly different at the 5% level
Table 5: Ethyl carbamate concentration in commercial wines obtained from the 1999, 2000 & 2001 Veritas Shows
* Total number of wines analysed per year, not total of cultivars illustrated ** ppb = parts per billion or g/L
Table 6: Urea formation (mg/L) by five yeasts in two wines from each of three different cultivars [CLICK IMAGE FOR LARGER VERSION]

Wynboer is incorporated in WineLand, magazine of the SA wine producers.

Subscribe to WineLand

Visit our sister sites:

South African wine farmers' representative organisation

Facts, figures, contact details and much more in the 2009/10 Directory

UP COPYRIGHT (C) 2000 WineLand