View Report Details

Assessment of, relative to other pathways, the contribution made by the food chain to the problem of quinolone resistance in microorganisms causing human infections

Project Code: B10004

31/03/2005

Veterinary Laboratories Agency, Weybridge
Snary, E

The emergence of antimicrobial-resistant microorganisms in both humans and food animals is a growing concern. For over 30 years, debate has centred on links between antimicrobial use in the production of food animals and the emergence of resistance in the human population. One particular class of antimicrobials that is under examination is the quinolones, which is a powerful class of antimicrobials that are used for therapeutic purposes in both human and veterinary medicine. Campylobacter is the most common cause of food-borne illness in the UK and since the introduction of quinolones in human and veterinary medicine, Campylobacter, in both humans and animals, has been quick to develop resistance. However the role of the food-chain (and in particular use of quinolones as a therapeutic in food-animal production) in the emergence of quinolone-resistant Campylobacter (QRC) in humans is uncertain. Consequently, the aim of this project is to assess, relative to other pathways, the contribution made by the food chain to QRC infection in humans. To achieve this, a quantitative risk assessment approach was adopted.

In order to address the risk question posed, a number of identified risk pathways were investigated. For convenience, each of which were allocated to one of three types (Food, Environment and Human). Those considered were:

Food: chicken (conventional, free-range, organic, non-UK); pigs; vegetable crops
Environment: public water supplies; pets (dogs and cats)
Human: foreign travel; human use of the ciprofloxacin for the treatment of campylobacter

The model considered a number of other potential pathways but due to data gaps/deficiencies a quantitative risk model could not be developed. These were: turkey, halal chicken, private water supplies, rendering, recreational water activities, wildlife, visits to pasture, and farm/abattoir workers. A qualitative review of each of these pathways includes a suggested model framework, available data to parameterise such a model and the identified data gaps/deficiencies.

Considering the pathways for which quantitative risk assessment was developed, it was concluded that the majority of QRC cases in the UK (59.6%) are attributable to QRC exposure while abroad. The 5th and 95th percentiles are 54.6% and 64.6% respectively). This is because, of an estimated average of 90,426 (37,919; 179,607) QRC cases per year, 52,801 (21,999; 105,930) were estimated to be associated with travel abroad.

Chicken consumption was predicted to be the next common source of QRC illness as it was estimated that chicken consumption was the source of approximately 8,000 QRC cases (8% of average total). Of these 1,139 (209; 3,034) and 6,946 (331; 24,183), or 1.3% and 6.7% of total QRC illnesses, were attributable to domestic and non-UK chicken consumption respectively. However, there is large uncertainty over the usage rates of domestic fluoroquinolone (FQ) use in chicken flocks, and particularly non-UK FQ usage rates. If FQ usage rates in non-UK chicken production are similar to the estimated domestic FQ usage rates, then the current estimate for the number of QRC cases per year attributable to non-UK chicken consumption is a vast over-estimate.

The contribution of free-range and organic chicken to the number of QRC cases per year was negligible. It was concluded that, on average, all cases attributable to domestic chicken consumption consumed chicken produced from conventional flocks treated with a quinolone whilst being reared. However, due to data deficiencies this estimate has a high degree of uncertainty associated with it and therefore, in comparison to the pathways quantitatively assessed, could be ranked 2nd, 3rd, 4th or 5th.

Other potentially important pathways include pets (dogs and cats) and clinical treatment of QSC. It is estimated that 2,724 (511, 6,798), or 3.63 (0.4;11.8) %, of all QRC cases are attributable to pet contact and that 285 (92; 527) QRC cases are attributable to clinical treatment of quinolone-susceptible Campylobacter (less than 1.5% of total cases). Again, as above, due to large uncertainties in the model results, both the pet pathway and clinical treatment could be ranked either 2nd, 3rd, 4th or 5th.

All other pathways quantitatively modelled (pig meat consumption, public water consumption, crop consumption) on average contribute less than one QRC case per year. It is therefore concluded that these sources of QRC are not significant for human QRC illness.

Approximately 63,895, or 70.7% of total cases (although this figure is uncertain) are estimated to be attributable to those pathways quantitatively modelled. Therefore, approximately 30% of QRC cases are attributable to other sources of QRC than those considered here. A proportion of these cases may have been from pathways, which we were not able to model due to data gaps and deficiencies (turkey, halal chicken, private water supplies, water sports, farm/countryside visits and farm/abattoir workers) and those that were outside the scope of the project (e.g. meat derived from sheep or cattle).

It has been estimated that illness caused by quinolone-resistant Campylobacter rather than a quinolone-susceptible strain results in 90,753 (36,594; 181,992) extra days of illness per year. No discernible difference in severity can be found at present. Approximately 84 (49; 123) people will suffer treatment failure because of illness with a quinolone-resistant strain of Campylobacter.

A number of the control strategies were investigated, including banning the use of quinolones in UK broiler production. Under this scenario, it was predicted that no QRC cases would be attributable to UK chicken (a reduction of 1,139 (209; 3,034)). Hence there would be a 1.3% decrease (by the mean) in the total number of human QRC cases This would reduce the number of cases failing FQ treatment by, on average, 1 (0; 4) per year. Diverting flocks treated with a fluoroquinolone to frozen chicken production would also reduce the number of QRC cases by 1,139. The debate on whether to cease the use of fluoroquinolones is still ongoing, but it is important to note that in food-animal production fluoroquinolones are prescribed for therapeutic purposes. This risk assessment has only investigated the consequence of fluoroquinolone use in broiler flocks; however there are many more issues that would also need to be considered to inform a risk management decision.

An alternative control strategy would be to reduce carcass contamination levels of Campylobacter by approximately 1 log at the end of processing. Here it was estimated that the number of Campylobacter and QRC cases attributable to chicken consumption would be reduced by approximately 100-fold, hence the average number of QRC cases would be reduced by approximately 1,130. Hence, control measures for campylobacter are also an effective way of reducing the number of human QRC cases (in addition to reducing the total number of Campylobacter spp. cases in humans).

The risk assessment identified a number of important data gaps/deficiencies within the quantified pathways, and we have recommended which model parameters should have priority for further investigation. Such data gaps include the issue of human immunity to Campylobacter, variability in pathogenicity to humans, survival of QRC in comparison to Campylobacter spp. and use of FQs in poultry production. A lack of validated, standardised breakpoints is also an issue. Data deficiencies include estimation of processing, preparation and consumption parameters within the food chain models including differentiating between eating at home and in restaurants or other catering establishments, dose-response models and the Campylobacter under-reporting factor. As a result of these data deficiencies, a number of the model outputs are uncertain and therefore the ranking of pathways was difficult. Although data limitation and data deficiencies are often seen as a weakness of risk assessment, the identifications of such data issues is an important output. In particular, through highlighting such data limitations or deficiencies, risk assessment can therefore help to prioritise future research in the area of QRC. It is therefore hoped that once such data has been collected the iterative nature of risk assessment will be utilised by updating the model with new data to produce an updated and more-certain estimate.

In conclusion, in answer to the question posed by the Food Standards Agency, it is estimated that 8% (mean value) of QRC cases in humans in the UK are attributable to the food chain (chicken, pig-meat and crops) and 1.3% to the consumption of domestic chicken. There is large uncertainty about this value. All cases attributable to consumption of UK produced food were attributable to chicken that were prescribed a fluoroquinolone (due to ill-health in the flock). The largest cause of QRC illness in humans was foreign travel as 52,801 (21,999; 105,930) cases were estimated to be attributed to this pathway, which is 59.6 (54.6; 64.6)% of the total number. Interpretation of all of the model results must be done with care because many data gaps/deficiencies were identified (and hence assumptions imposed) during the development of the risk assessment. Further, targeted, research could reduce such uncertainties.

There are no documents uploaded for this report.

Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: