Descriptive study and monitoring of the symptomatic signs reported by consumers of the drinks marketed under the brand name Coca-Cola in June 1999

Introduction

Following the notification by the Belgian authorities of health problems reported by consumers of Coca-Cola drinks produced in Antwerp, Ghent and Dunkerque, the sale of these drinks was suspended in France on15 June 1999 and the Centres Anti-Poisons (CAP - Anti-Poisoning Centres) were put on the alert.

From that date, the CAPs and the Customer Service department of Coca-Cola began receiving a large influx of calls reporting health problems. Microbiological and toxicological analyses as well as the expert appraisal of the production processes and of the quality control procedures eliminated any possibility of the contamination having taken place during the production or the packaging processes. Nevertheless, a degree of uncertainty remained concerning the risk of contamination during the storage and distribution phases.

Consequently, the DGS, the CAPs and the InVS set up an epidemiological surveillance process with the aim of

1) describing the phenomenon,

2) observing the return to normal,

3) detecting the appearance of sets of cases with a view to possible investigations.

Method

An "alleged" case was defined as any person having reported either to the CAPs or to the Coca-Cola Consumer Department, any health problems following the ingestion of Coca-Cola brand products as of 14 June 1999. Information concerning the complaints about the quality of the products as well as consumption and distribution data was collected by the company Coca-Cola. A first analysis was carried out for the period 14 June to 5 July.

Results

In total, 1930 cases were reported, half of which between the 16th and 17th of June. The most frequent symptoms reported were abdominal pain (54%), diarrhoeas (34%) and head aches (30%). The time lapse between the last ingestion and the appearance of the symptoms varied between one hour and several days. The number of reported cases per million litres distributed was higher in the North of France (9.26) and lower in the South West and South East (respectively 3.7 and 3.2). The frequency of cases varied with age (p< 10-5); the highest concerned consumers aged between 30 and 39 years and the lowest was among the 12 to 19 year-olds. For the latter, the frequency of the cases appeared to be higher among males (p = 0.08).

The results of the study summarised above have been the subject of a presentation made by Martine Ledrans, an epidemiologist with the Health and Environment Department of the Institut de Veille Sanitaire (Health Watch Institute), on the occasion of the General Assembly of the European Institute of Cindynics on 2 March 2000.

During and following this presentation, the following points were discussed concerning the origin of this epidemiological phenomenon:

- In all of the regions, a peak in the number of reported cases was observed between the16th and 18th of June, that is the 3 days following the press release announcing the suspension of the distribution of the Coca-Cola drinks prepared in Dunkerque. A maximum of 559 cases was reported on 17th June. Starting from July 1st, less than 10 cases were being reported every day.

- The results obtained do not allow one to conclude in favour of a toxic or infectious aetiology common to all cases. Indeed:

- The symptoms reported are very diverse. Certain reported cases cannot be reasonably imputed to the consumption of the drink, even were it to be shown that the drink had indeed been chemically or microbiologically contaminated.

This is particularly true for the cases reporting serious or chronic ailments (renal insufficiency, intestinal haemorrhage, rectorragies). It should be noted that considering the very high number of consumers of Coca-Cola in France (42% of the population), one may expect to observe serious health conditions outside of any relationship with the quality of the drink absorbed. Besides, it is difficult to identify a homogeneous set of patients sufficient to establish a typical clinical table. The extremely variable incubation time according to the cases also illustrates this heterogeneousness.

- The cans manufactured in Dunkerque are distributed for sale in the Northern half of France. Yet, a spatial analysis revealed an epidemic phenomenon that was similar throughout France, even though it was less widespread in the Southern half.

- The higher frequency of cases was observed in the Northern commercial region, where the plant from which the suspended drinks was located. Moreover, during the event, several press releases originating from the CAP in Lille reinforced the local media effect. One should also note the proximity with the Belgian border, where the fact of sets of cases affecting schools received high media attention. The media attention given to the event in Belgium may also explain the fact that the Eastern region of France came second in the number of cases.

- The estimated incidence of cases was higher among consumers of Coca-Cola and Coca Light than for other products. This could be explained by the influence of the information messages and the headlines, which targeted the Coca-Cola brand rather than the individual products listed in the press release suspending the distribution of the products.

- A study similar to ours was carried out by the CAP in Brussels following calls received between the 8th and the 20th of June. Figure 2 shows that distribution of symptoms was different between the Belgian and the French cases, with a predominance of headaches and vomiting in the first instance and abdominal pain and diarrhoea for the second. One therefore notes a certain discordance between the data.

- Two explanations may be given to explain all or part of the phenomenon observed:

- People might have attributed the occurrence of symptoms having an organic cause to the consumption of drinks coming from the Coca-Cola brand.

This hypothesis is plausible, considering the frequency and the low specificity of the symptoms reported and the consumption habits of the population. One should remember that 42% of French people are regular or occasional consumers of the Coca-Cola drink alone.

- The anxiety of the consumers alerted by the media or their entourage as to the suspicion which weighed on the products coming from the Coca-Cola brand might in and of itself have been a cause of the manifestations reported among consumers. This explanation was suggested for the Belgian phenomenon by B. Nemery who noted that, "several characteristics of this epidemic would tend to indicate a phenomenon of 'mass sociogenic illness (M.S.I.)'" (1). MSI is described in the literature as a constellation of multiple symptoms of an organic disease having no identifiable cause and happening with two or more people who share the same convictions as to the origin of these symptoms (2). Several authors have nevertheless debated Nemery's interpretation (3 , 4 ) in particular for the first episode affecting a Belgian school, where the Coca-Cola consumed from a bottle presented a strong smell and a bad taste due to the presence of sulphured hydrogen. Alerts concerning the quality of certain food products had occurred a few weeks earlier both in Belgium and in France (dioxins in chicken, contamination of a certain bottled water for retail distribution). The media coverage of his first episode, in such a "favourable" climate, could have lead to the hatching of MSI phenomena, in communities that are particularly sensitive to MSI phenomena, such as schools (2) and those phenomena, which are more rarely described in the literature on the scale of a whole country such as France.

Martine Ledrans, Henriette De Valk, Philippe Germonneau, Quénel Philippe, Institut de Veille Sanitaire, (InVS)

Boudot Jocelyne Direction Générale de la Santé. (DGS)

Ilef Danièle Cellule Interrégionale d'Epidémiologie du Nord Pas de Calais

Grandbastien Bruno Hôpital Calmette de Lille

Garnier Robert Centre Anti-Poisons de Paris

Manel Jacques Centre Anti-Poisons de Nancy

Monique Mathieu Nolf Centre Anti -Poisons de Lille , Pulce Corinne Centre Anti-Poisons de Lyon.

Translation by Andrew WILES

Bibliography :

1. Nemery B. et al. Dioxins, Coca-Cola, and mass sociogenic illness in Belgium. Lancet 1999 ; 354.

2. Boss L. P Epidemic Hysteria : a review of the Published Literature. Epidemiologic reviews. 1997 ; 19 : 2 : 233-242.

3. Van loock F., Gallay A., Demarest S., Van Der Heyden J ., Van Oyen H. Outbreak of Coca-Cola related illness in Belgium : a true association. Lancet 1999 ; 354 : 680-681.

4. Pouthier D. (letter to the editor) Lancet 1999 ; 354 : 681-682.