By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
A. Kaesbohrer. Federal Institute for Risk Assessment, Department for Biological Safety, National Reference Laboratory for Antimicrobial Resistance, Diedersdorfer Weg 1, D-12277 Berlin, Germany. Tel.: +49 30 18412 2202; Fax: +49 30 18412 2952; E-mail: email@example.com
In 2009, 1462 Escherichia coli isolates were collected in a systematic resistance monitoring approach from primary production, slaughterhouses and at retail and evaluated on the basis of epidemiological cut-off values. Besides resistance to antimicrobial classes that have been extensively used for a long time (e.g. sulphonamides and tetracyclines), resistance to (fluoro)quinolones and third-generation cephalosporins was observed. While in the poultry production chain the majority (60%) of isolates from laying hens was susceptible to all antimicrobials tested, most isolates from broilers, chicken meat and turkey meat showed resistance to at least one (85–93%) but frequently even to several antimicrobial classes (73–84%). In the cattle and pig production chain, the share of isolates showing resistance to at least one antimicrobial was lowest (16%) in dairy cows, whereas resistance to at least one antimicrobial ranged between 43% and 73% in veal calves, veal and pork. Resistance rates to ciprofloxacin and nalidixic acid in isolates from broilers were 41.1% and 43.1%, respectively. Likewise, high resistance rates to (fluoro)quinolones were observed in isolates from chicken meat and turkey meat. In contrast, ciprofloxacin resistance was less frequent in E. coli isolates from the cattle and pig production chain with highest rate in veal calves (13.3%). Highest resistance rates to cephalosporins were observed in broilers and chicken meat, with 5.9% and 6.2% of the isolates showing resistance. In dairy cattle and veal, no isolates with cephalosporin resistance were detected, whereas 3.3% of the isolates from veal calves showed resistance to ceftazidime. Resistance to (fluoro)quinolones and cephalosporins in E. coli isolates is of special concern because they are critically important antimicrobials in human antimicrobial therapy. The emergence of this resistance warrants increased monitoring. Together with continuous monitoring of antimicrobial usage, management strategies should be regularly assessed and adapted.
• Continuous monitoring concept implemented A systematic resistance monitoring was established in Germany involving official veterinary services and covering all major food production chains within a 3-year interval to describe the magnitude of the problem.
• Commensal bacteria as indicator for risks All animal species used for food production as well as humans carry E. coli in their intestinal tract. This is a good indicator for the diversity of resistance information in the population.
• Antimicrobial resistance in commensal E. coli reflects the risk for consumers Resistance to antimicrobials in zoonotic bacteria and commensals may compromise the effective treatment of human infections. Resistance to antimicrobials which are considered critically important for human therapy was observed. Management strategies need to be regularly adapted to limit these risks.
Antimicrobial agents are essential for the prevention, control and treatment of bacterial infections in humans and animals (Aarestrup, 1999; FAO/WHO/OIE, 2008). The use of antimicrobial agents, however, can cause the emergence, prevalence and dissemination of antimicrobial resistance in target pathogens and normal bacterial flora. Bacterial resistance to antibiotics has increased in recent years worldwide (EFSA, 2009). Resistance to antimicrobials in zoonotic bacteria and commensals may compromise the effective treatment of infections in humans. The WHO list of critically important antimicrobials for human medicine includes third- and fourth-generation cephalosporins and fluoroquinolones. The effectiveness of critically important antimicrobials for human medicine should not be compromised by inappropriate overuse and/or misuse in the agricultural sector (FAO/WHO/OIE, 2008). During the last decade, a number of events have stressed the need for increased awareness of public health aspects related to antimicrobial resistance in animal husbandry. For example, the emergence of beta-lactamases shows that risks to human health include the possibility of horizontal transfer of resistance genes. Thus, foodborne risks go beyond spread of zoonotic bacteria such as Salmonella and Campylobacter (CVMP, 2011).
Resistant foodborne pathogens and commensals can be transmitted to humans via several pathways (FDA, 2001; EFSA, 2007, 2008c; FAO/WHO/OIE, 2008). Exposure of the public through consumption and handling of contaminated food, direct contact with the animals and release of resistant bacteria into the environment may contribute to the spread of resistance determinants. The need was highlighted to have in place antimicrobial resistance surveillance for animals, humans and food (EFSA, 2008c, 2009; FAO/WHO/OIE, 2008). Risk analysis shall be performed for assessing the overall risk to human health from foodborne antimicrobial-resistant microorganisms and determining appropriate risk mitigation strategies to control those risks following Codex Alimentarius principles (Codex Alimentarius, 2011).
In the European Union, a monitoring concept was drafted to harmonize implementation of Directive 2003/99/EC (Kaesbohrer and Heckenbach, 2006). Detailed technical specifications for harmonized monitoring of antimicrobial resistance of Salmonella and Campylobacter in Europe were published by EFSA (2008a). To overcome limitations in availability of uniform data on antimicrobial resistance in commensal indicator organisms, EFSA also published a guideline for harmonized monitoring and reporting of antimicrobial resistance in Escherichia coli and Enterococcus spp. from healthy farm animals (EFSA, 2008b). Because of their widespread availability, monitoring of commensal bacteria allows comparison of the effects of selective pressure in all relevant populations and is considered useful as an early alert system, for tracking emerging resistance in livestock and possible spread to animal-derived food (EFSA, 2008b).
Minimum inhibitory concentration (MIC) break points used for predicting clinical efficacy of antimicrobials might differ from those applied for monitoring purposes. Because the monitoring approach should aim at optimum sensitivity for detection of acquired resistance, epidemiological cut-off values are proposed as interpretive criteria and not clinical break points (Aarestrup et al., 2007; EFSA, 2008b). These criteria have been developed by the European Committee for Antimicrobial Susceptibility Testing (http://www.eucast.org).
The overall objective of the national monitoring programme described in this study is to estimate the prevalence of resistance to antimicrobial substances relevant for public health along the food chain and to identify factors with major impact on the related risks. The continuation of the present resistance monitoring activities will allow the assessment of changes in the prevalence rate of antimicrobial resistance.
Materials and Methods
A systematic resistance monitoring approach was established in 2009 in Germany covering each of the major food production chains within a 3-year interval at least once. During the first year of implementation, monitoring covered laying hens for table egg production, broilers of Gallus gallus as well as chicken meat and turkey meat from the poultry production chain, dairy cattle, veal calves and veal from the cattle production chain and pork from the pig production chain. To meet the objectives, all isolates from a representative sample from each of the targeted populations are tested at least once within the 3-year period. To estimate the prevalence of zoonotic agents, 384 specimens per year, bacterial species and population considered are investigated. The calculation is made for a 50% prevalence, an accuracy of 5% and a 95% confidence level. According to Commission Decision 2007/407/EC and the EFSA recommendation (EFSA, 2008b), the target number for antimicrobial susceptibility testing is 170 for each bacterial species, although it is anticipated that this may not be feasible for each study population. Faecal samples or food samples are collected within this active monitoring programme, which is run at farm, abattoir and retail level by the regional authorities. Samples are used to estimate both the prevalence of the micro-organism at the specific level and antimicrobial resistance among the isolates collected. Bacterial agents covered in the monitoring include the most relevant zoonotic pathogens (e.g. Salmonella, Campylobacter, verotoxigenic E. coli) as well as E. coli as indicator organisms. For each of the populations, a stratified sampling plan was designed. Stratification allocated to each Federal State a subset of sampling units. Allocation of samples was meant to be proportional to the respective animal population for primary production samples, to the respective slaughterhouse capacity for samples at the abattoir and to the human population for samples at retail. For laying hens and broilers, samples taken in the context of the salmonella eradication programme were used for collecting E. coli isolates. For each of the units, one isolate was selected for resistance testing to ensure independency of results. All samples were collected by official veterinarians taking into account specific instructions for sampling, storage and transport of samples to official laboratories. For collection of E. coli isolates, faecal samples were plated directly on media selective for E. coli, for example, MacConkey agar. Isolation of E. coli from bulk tank milk samples of dairy herds and meat samples collected at retail was carried out according to ISO 16649.
Resistance of isolates to 13 antimicrobials (gentamicin, kanamycin, streptomycin, chloramphenicol, florfenicol, ampicillin, cefotaxime, ceftazidime, nalidixic acid, ciprofloxacin, sulphamethoxazole, trimethoprim and tetracycline) was investigated using the broth microdilution method according to NCCLS/CLSI standards M31-A3 (2007) at the National Reference Laboratory for Antimicrobial Resistance. Throughout this study, resistance against more than one antimicrobial class was called multidrug resistance. Furthermore, cephalosporins (cefotaxime, ceftazidime) and other beta-lactams (ampicillin) were considered separate classes. Minimum inhibitory concentrations were evaluated according to epidemiological cut-off values (http://www.eucast.org, accessed 17 May 2010). Cut-off values for the antimicrobials (in μg/ml) used throughout the study were as follows: gentamicin (2), kanamycin (8), streptomycin (16), chloramphenicol (16), florfenicol (16), ampicillin (8), cefotaxime (0.25), ceftazidime (0.5), nalidixic acid (16), ciprofloxacin (0.03), sulphamethoxazole (256), trimethoprim (2) and tetracycline (8). Isolates with an MIC above the cut-offs were considered as resistant, otherwise as susceptible. Cut-off values used were in line with EFSA recommendations (EFSA, 2008b), whereas the range tested was specific for Germany.
Altogether, 1462 E. coli isolates were collected within the programmes of the first year, ranging from primary production to meat at retail. The number of available isolates varied between the programmes. While in primary production and at slaughter analysis of faecal samples resulted in at least 170 E. coli isolates, in meat samples, this number could only be achieved for chicken and turkey meat, but not for milk from dairy cows (93 isolates), veal (51 isolates) and pork (46 isolates).
Results from antimicrobial susceptibility testing of E. coli isolates in poultry and poultry products are shown in Table 1. In the poultry production chain, the majority of isolates (59.6%) from laying hens was susceptible to all antimicrobials tested, whereas the proportion of fully susceptible strains ranged between 6.9% for turkey meat and 15.3% for broilers. Most isolates from broilers, chicken meat and turkey meat were resistant to at least one but frequently to several antimicrobial classes. Rates for resistance to more than one class were 23.4% (laying hens), 72.8% (broilers), 73.7% (chicken meat) and 84.2% (turkey meat) and to more than four classes were 3.8%, 17.8%, 19.6% and 38.4%. In Fig. 1, resistance rates for tetracycline, ciprofloxacin and ceftazidime in the study populations are compared. In broilers and chicken meat, highest resistance rates were observed against sulfamethoxazole and ampicillin. In turkey meat, resistance against tetracyclines was higher compared to chicken meat (82.8% versus 45.4%). Resistance rates to ciprofloxacin and nalidixic acid were higher in isolates from broilers (43.1% and 41.1%, respectively) and chicken meat (53.1% and 50.5%) compared to turkey meat (30.0% versus 26.1%). Resistance to third-generation cephalosporins was mainly detected in broilers and chicken meat (5.9% and 6.2% for ceftazidime, respectively) and lower in laying hens and turkey meat (1.3% and 1.0% for ceftazidime, respectively).
Table 1. Number and proportion of resistant Escherichia coli isolates from the poultry production chain
Source of isolates
Laying hens N (%)
Broilers N (%)
Chicken meat N (%)
Turkey meat N (%)
Number of isolates
For cattle, resistance rates in isolates from dairy cattle, veal calves and veal differed considerably (Table 2). While only 16.1% of isolates from milk cows showed resistance against at least one antimicrobial class, isolates from veal calves and veal were resistant in 72.9% and 62.7% of cases to one or several antimicrobial classes. Rates for resistance to more than one class were 5.3% (dairy cattle), 69.0% (veal calves) and 50.9% (veal) and to more than four classes were 2.2%, 22.2% and 9.8%. Similar to the situation in isolates from poultry and poultry meat, resistance to ampicillin, sulfamethoxazole, streptomycin and tetracycline predominated in bovine isolates. Resistance to third-generation cephalosporins was detected in veal calf isolates: 1.4% for cefotaxime and 3.0% for ceftazidime. None of the 93 isolates from dairy cattle and 51 isolates from veal was resistant to third-generation cephalosporins. In bovine isolates, resistance to (fluoro)quinolones was highest in veal calves (13.3% for ciprofloxacin) compared to 2.2% and 3.9% for dairy cattle and veal, respectively.
Table 2. Number and proportion of resistant Escherichia coli isolates from the cattle and pig production chain
Source of isolates
Dairy cattle N (%)
Veal calf N (%)
Veal N (%)
Pork N (%)
Number of isolates
Among the 46 isolates from pork, 43.5% were resistant to at least one antimicrobial (Table 2). Rates for multidrug resistance were 32.6% (more than one class) and 8.7% (more than four classes). Similar to the situation in poultry and cattle, resistance to ampicillin, sulfamethoxazole, streptomycin and tetracycline predominated. One (2.2%) of 46 pork isolates was resistant to third-generation cephalosporins, and three isolates (6.5%) were resistant to nalidixic acid and ciprofloxacin, respectively.
Escherichia coli is one of the most important foodborne pathogens. Infections caused by antimicrobial-resistant E. coli are becoming increasingly common worldwide and pose a serious health problem for human medicine (EARSS, 2008; ECDC, 2010). In most countries reporting data to EARSS and its successor EARS-Net, a significant increase in resistance to third-generation cephalosporins was observed from 2005 to 2008. Furthermore, resistance to fluoroquinolones in E. coli from bloodstream infections has increased all over Europe consistently over recent years (ECDC, 2010).
The degree to which animals and humans share or exchange common types of E. coli is currently the subject of active research and debate (EFSA, 2011). Commensal E. coli, present in the intestine of farm animals or on animal products, are considered a potential source of resistant bacteria and resistance genes for humans. The German resistance monitoring programme provides valuable data on the pool of resistance present in bacteria of animal origin. This will contribute to the assessment of their relevance for public health.
The results generated within this first year of active monitoring of resistance in commensal E. coli in Germany reflect that these bacteria show resistance to various antibiotics relevant to human medicine. The results are in line with the resistance situation in the food chain previously described for Germany on the basis of Salmonella isolates submitted to the National Reference Laboratory for Salmonella at the Federal Institute for Risk Assessment between 2000 and 2008. Like in the present study, those data had been analysed on the basis of epidemiological cut-off values (Schroeter and Kaesbohrer, 2011).
A previous German study (1999–2001) on 317 E. coli strains from cattle, swine and poultry showed that resistance was found in 40% and multidrug resistance (2–8 resistance determinants) in 32% of the strains (Guerra et al., 2003). Reported resistance was significantly higher in isolates from poultry (61%) and swine (60%) than from cattle (25%) at that time. Resistance to sulfamethoxazole, tetracycline, streptomycin, ampicillin and spectinomycin (30–15%) predominated. Eleven per cent of the strains were resistant to nalidixic acid. None of the isolates showed resistance to ceftiofur. As clinical break points were applied throughout that study, data are not directly comparable with current results, but relations between different sources of isolates can be compared.
On EU level, in 2009, the occurrence of resistance to tetracyclines, ampicillin, chloramphenicol, streptomycin, ciprofloxacin and nalidixic acid in E. coli isolates from chicken (without detailed specification of the production line) showed considerable variation between reporting Member States, similar to the situation observed in previous years (EFSA, 2011). In a pan-European survey of antimicrobial susceptibility in E. coli from broilers sampled at slaughterhouse carried out in 2002/2003, De Jong et al. (2009) described high resistance rates to ampicillin, chloramphenicol and tetracycline on the basis of clinical break points. The resistance prevalence was higher in older compounds (e.g. ampicillin, tetracycline and trimethoprim/sulfamethoxazole) compared to the newer compounds, for example, cefotaxime and ciprofloxacin (De Jong et al., 2009). The present study confirms this tendency that resistance to sulfamethoxazole, tetracycline, streptomycin and ampicillin is frequent in poultry isolates. Furthermore, highest resistance rates to ciprofloxacin were reported in isolates from broilers (43.1%) and chicken meat (53.1%). This is in line with Dutch (MARAN, 2009) and EU-data (EFSA, 2011). In the Netherlands, resistance rates tended to increase in isolates from broilers and broiler products over time (MARAN, 2009). The respective values for ciprofloxacin in the older pan-European study were 30.4% decreased susceptibility and 5.8% clinical resistance in isolates from broilers (De Jong et al., 2009). Decreased susceptibility to fluoroquinolones has been associated with decreased clinical responses of Salmonella infections to fluoroquinolones in humans (Crump et al., 2003; Helms et al., 2004).
Likewise, in this study, broiler and chicken meat isolates showed the highest resistance rates to cefotaxime (5.4% and 6.2%, respectively). The main cause of cefotaxime resistance is the production of different extended-spectrum beta-lactamases (ESBL) or AmpC beta-lactamases as previously described for Salmonella from animals and food products (Rodriguez et al., 2009). In the pan-European study, similar results were reported for broilers. Whereas 5.4% of the isolates showed decreased susceptibility for cefotaxime on the basis of epidemiological cut-off values, 0.4% were clinically resistant on the basis of the clinical break point according to CLSI (De Jong et al., 2009). In the report on EU level, resistance in E. coli to cefotaxime in chicken ranged in reporting countries from 2% to 26% (EFSA, 2011). In the Netherlands, in E. coli isolates from poultry meat, 21.3% and 18.0% resistance to cefotaxime and ceftazidime was observed (MARAN, 2009).
Within our study population, considerable rates of multidrug resistance (resistance to more than one antimicrobial class) were observed among isolates from broilers, as well as chicken and turkey meat. Highest multidrug resistance rates were observed in turkey meat with 38.4% of the isolates showing resistance to more than four classes. Similarly, in the Netherlands, 27.5% of the isolates from broilers showed resistance to six or more classes of antimicrobials.
In Germany, sales data for antimicrobials used in veterinary medicine are only available for the year 2005. Sales and consumption data on antimicrobials specific for each individual animal species are not available. In the Netherlands, it is reported that the average broiler chicken receives 37 daily doses per year, which means that an individual broiler is treated with antibiotics during 4 days in the 42 days from hatching until slaughter (MARAN, 2009). Quinolones accounted for 18% of total antibiotic use in broiler farms, which was only exceeded by the usage of penicillin. Fluoroquinolone use was 1.4% of total use. This may explain to some extent the resistance rates observed. In a recent review, the relationship between antimicrobial usage and prevalence of antimicrobial-resistant bacteria from food-producing animals was confirmed in Japan (Harada and Asai, 2010). It was concluded that trends in antimicrobial resistance are closely related to the use of antimicrobial agents in veterinary medicine (Aarestrup, 1999; Asai et al. 2005; Harada and Asai, 2010). In contrast, cephalosporins are not licensed for usage in poultry. Therefore, the reason for the presence of resistance to cephalosporins in poultry, as observed in this and several other recent studies, is unclear. For example, in a study on the occurrence of ESBL-producing E. coli in flocks of laying hens reared in Danish organic systems, the factors leading to the origin and persistence of cephalosporin resistance remained unknown (Bortolaia et al., 2010a). In the Netherlands, resistance to cephalosporins could be linked to usage in hatcheries. Similarly, in Quebec, changes of ceftiofur resistance in chicken Salmonella Heidelberg and E. coli isolates were clearly related to changing levels of ceftiofur use in hatcheries (Dutil et al., 2010).
The levels of resistance to all of the antimicrobials in isolates of E. coli from cattle were consistently lower compared to the levels in indicator E. coli from Gallus gallus and pigs from the same Member States (Guerra et al., 2003; EFSA, 2011; MARAN, 2009; De Jong et al., 2009). Detailed analysis of data collected in the present study confirms that resistance level in dairy cattle is low, whereas resistance level in veal calves is higher (MARAN, 2009; De Jong et al., 2009). In Germany, 5 (1.4%) of the E. coli isolates from veal calves were resistant against cefotaxime, whereas 11 isolates (3.0%) were resistant to ceftazidime. In the Netherlands, resistance in veal calves to third-generation cephalosporins was comparable to the German situation, ranging between 1.8% and 2.3% (MARAN, 2009).
Differences in resistance to quinolones between veal calves and beef cattle were also observed. In the Netherlands, 18.1% of isolates from veal calves showed reduced susceptibility, and 6.5% were considered clinically resistant (MIC > 1 mg/l) against ciprofloxacin. In the present study, 13.3% of the isolates from veal calves showed resistance to ciprofloxacin. In Italy, 5.3% of the cattle isolates showed clinical resistance against ciprofloxacin. In contrast, during the EASSA study in 2002/2003, lower resistance rates for fluoroquinolones had been reported for beef cattle, and no clinical resistance against ciprofloxacin was observed for fluoroquinolones in four of five countries (De Jong et al., 2009).
Among the bovine isolates, multidrug resistance was most frequent in veal calf isolates with 22.2%, showing resistance to more than four classes. Similarly, among E. coli collected in the Netherlands from veal calves, multidrug resistance was also widespread with 11.1% of the isolates resistant to six or more classes of antimicrobials (MARAN, 2009). In contrast, within our study, 2.2% of isolates from dairy cattle showed resistance to more than four classes. Differences observed may be correlated with differences in usage patterns. Whereas in dairy cattle, all antimicrobials are applied parenteral or intramammary, oral use is the typical route of application in calves (Merle et al., 2011).
In the EFSA report 2011, the resistance rate of E. coli from pork to tetracyclines was 41% for all reporting countries. The corresponding figures for ampicillin were 27%, for chloramphenicol 6%, for sulphonamides 35% and for streptomycin 41% (EFSA, 2011). In Germany, the observed level of resistance for all these antimicrobials was lower compared to the EU average, except for chloramphenicol, where similar rates were detected. Rates of resistance in E. coli from pork to nalidixic acid remained at low or moderate levels in all reporting countries throughout 2005–2009. Our results from Germany are comparable to EU average data for ciprofloxacin and nalidixic acid, where resistance rates in E. coli isolates from pork were 6% (both antimicrobials).
Resistance to cephalosporins seems to increase in E. coli from pork. In 2009, all countries reporting to EFSA detected cefotaxime resistance in E. coli isolates from pork. The EU average rate for cefotaxime resistance was 3%. In the previous year, only France reported cefotaxime resistance in one of 102 E. coli isolates from pork (EFSA, 2011).
Overall, 68% of the 1462 commensal E. coli strains isolated in the present study were resistant to at least one antimicrobial, and 57% were resistant to more than one antimicrobial class. Resistance was frequently detected against those antimicrobials frequently used in Germany in the veterinary field, for example, tetracyclines, aminopenicillins and sulphonamides (Merle et al., 2011). Whereas most isolates from laying hens were fully sensitive, the majority of isolates from poultry for fattening and from veal calves were resistant to several antimicrobials. Obvious differences observed in resistance rates between the animal production groups can be explained by differences in the amount of usage of antimicrobials for treatment. The level of resistance to fluoroquinolones (ciprofloxacin) in E. coli from poultry is alarming, where values ranged between 30% and 43%. Resistance to the third-generation cephalosporins was prevalent in E. coli from most populations tested, ranging from 1.0% in turkey meat to 6.2% resistant isolates in chicken meat.
Several international organizations encourage the establishment of surveillance studies on antimicrobial resistance and antimicrobial use surveillance worldwide. Therefore, harmonized reporting and use of harmonized break points within and between participating countries are necessary to ensure the comparability of data. To implement those systems, existing successful surveillance systems can be used as models. Special attention should be paid to antimicrobial agents regarded as critically important for use in humans.
We wish to thank the regional authorities and laboratories of the federal states for carrying out the monitoring and submitting the isolates. We would also like to thank them for the years of good cooperation. We wish to extend our particular thanks to the scientific and technical staff in laboratories at BfR also. Without their excellent support, this paper would not have been possible.
CLSI, Clinical and Laboratory Standards Institute; CVMP, Committee for Medicinal Products for Veterinary Use; DANMAP, Acronym for: Danish Integrated Antimicrobial Resistance Monitoring and Research Programme; EARSS, European Antimicrobial Resistance Surveillance System; EC, European Community; ECDC, European Centre for Disease Prevention and Control; EFSA, European Food Safety Authority; ESBL, Extended Spectrum Beta Lactamase; FAO, Food and Agriculture Organization of the United Nations; FDA, Food and Drug Administration; ISO, International Standards Organisation; MARAN, Monitoring of Antimicrobial Resistance and Antibiotic Usage in Animals in the Netherlands; MIC, Minimum Inhibition Concentration; NCCLS, National Committee on Clinical Laboratory Standards; OIE, World Organisation for Animal Health (Office International Epizootics); WHO, World Health Organisation.
Conflicts of Interest
All authors confirm that they have no conflicts of interest.