The 38 veterinarians from Attica who completed the questionnaire represent approximately 10% of the population of small animal practitioners in the area of Attica (source: HVA). This sample size is similar to that of other early studies conducted elsewhere [15, 16]. Participation was encouraging (response rate: 69%) and extending the study to the rest of Greece should be the next step, producing a more representative set of results for the whole country.
Almost half of the veterinarians answered that they have a written antibiotic policy in their practice. This compares to 27% in South Africa and 30% in Denmark [16, 17]. In the UK, 97% of practicing veterinarians responded in earlier studies that they did not use fixed protocols . Having written antibiotic protocols is important as it promotes optimal treatment, controlling at the same time the prescription of antimicrobials of critical clinical importance in human medicine, e.g. fluoroquinolones and cephalosporins of 3rd and 4th generation which should only be used following culture and susceptibility testing. Despite the apparent large proportion of respondents affirming that they have written protocols, it is important to state that there are no guidelines on the use of antibiotics by any veterinary authority in Greece. Each practitioner develops his own protocols, possibly leading to heterogeneity in the antibiotic treatment of clinical cases, and over-prescription of “last resort” antimicrobials when the practitioner is not up to date in the rational use of antibiotics. This likely heterogeneity is compounded by the lack of geographically explicit AMR surveillance data to inform practitioners’ protocols.
Of concern is that the majority of pet-owners administer antibiotics before they bring their animals to the practice. Pet owners are probably using antimicrobials without taking into consideration their spectrum of activity or resistance profiles, and they are either using remnants from previous visits to a veterinarian or even antibiotics used in human medicine. In Greece, there is legislation stipulating that antibiotics should only be prescribed by health professionals. Eurobarometer data on human antibiotic use for Greece in 2013 and 2016 showed that 16% and 20%, respectively, of all users do so without a medical prescription, the highest percentages in the EU . This figure decreased to 9% in 2018, indicating success in the implementation of policy recommendations as suggested by the European Commission. However, Eurobarometer data indicates that Greece is still one of the EU countries with the lowest awareness and understanding of antibiotic use by the general population .
Use of antimicrobial combinations is not frequent; our results are similar to those of a study in Italy , but in contrast with a study in the UK that showed that antibiotic combinations were commonly prescribed . We did not enquire further about the specific combinations; these should be investigated in a next extended survey. Antimicrobial combination is suggested to achieve synergistic or additive results, allowing lower doses of compounds and thus preventing emergence of antimicrobial resistance .
Almost every clinician reported owner-compliance challenges; this usually has to do with pet owners stopping treatment when clinical signs disappear (a general rule of thumb is to continue treatment for 1–2 days beyond resolution of clinical signs and for 2 weeks in serious deep pyoderma problems) . This should be further investigated in a future more extensive study, as it is important to educate pet owners accordingly.
Most of the veterinarians answered that they use microbiological analysis and antimicrobial susceptibility testing, alas with variable frequency; most (56%) answered “sometimes” and 41% answered “frequently”. A similar result was obtained by Barbarossa et al in Italy  (69% and 20%, respectively, while 2% answered “always”). However, only 27% of our respondents answered that they prefer to wait for the antibiogram results before using any antibiotics, and 94% answered that they have to use antimicrobials while waiting for the laboratory results, with variable frequency (from “sometimes” to “always”). These answers indicate that the first choice of antibiotic is usually made empirically; most answered that they based their choice on bibliography (e.g. therapy handbooks), which we assume is their main source for compiling their written antibiotic policies. The empirical use of antimicrobials should be avoided, with the exception of infections causing pain or discomfort or for complicated or life-threatening infections . There is an urgent need to reinforce the use of laboratory analysis in cases where multiresistant bacteria may be implicated or there is a risk of increasing antimicrobial resistance, e.g. Staphylococcus spp. and E. coli in cutaneous and urinary tract infections, respectively. For those cases where empirical use of antimicrobials cannot be avoided, up-to-date information on local resistance profiles would be critical to narrow down antibiotic choice.
Most veterinarians (85%) do not see animals returning to the practice because of antibiotic treatment failure. This may explain why 69% of respondents answered that they do not keep their clients’ record of antibiotics prescription, believing it is not important or due to lack of time. Although most respondents weigh the animals before prescribing antibiotics, answers regarding the types of antibiotics used and duration of treatment per organ system are concerning. Our results suggest that practitioners who are unaware of local antimicrobial resistance patterns or without laboratory results to consult would use top-shelf antimicrobials of broad-spectrum and suggest prolonged durations of treatment.
Forty seven percent of the veterinarians treat urinary tract infections with (fluoro)quinolones, usually for more than 8 days. Relevant guidelines suggest the use of (fluoro)quinolones only in upper urinary tract infection (pyelonephritis) in dogs and cats; most cases of cystitis should be treated with amoxicillin or trimethoprim/sulphonamides, optimally after laboratory results confirm the presence of bacteria and for a period of 3–5 days . The frequent isolation of ESBL-producing E. coli from the urinary tract of dogs and cats is of increased significance due to the risk of infecting owners and veterinarians. Thus, it is important to limit their advance with increased microbiological monitoring and rational antibiotic use.
Different treatment durations than the ones proposed by guidelines on rational use of antibiotics were reported for other organ systems as well. In GIT infections almost half of the respondents (48%) apply antimicrobials for 8–14 days, while the general recommendations are for 3–7 days . Moreover, metronidazole is used by 1/3 of the respondents as the first choice for GIT infections; metronidazole is critical in managing Clostridium difficile infections in humans, and countries like Denmark have managed to reduce its use in veterinary medicine to nearly zero levels after a campaign initiated in 2012 .
In the present survey, respondents answered that cephalosporins of 1st and 2nd generation, and penicillins are the preferred choices for dealing with skin infections, which is encouraging. Cutaneous infections are important in terms of antimicrobial resistance, with methicillin resistant strains of Staphylococcus aureus and Staphylococcus pseudintermedius presenting a serious threat to animal health due to the increased risk of treatment failure. These bacteria are recognised nosocomial pathogens and can spread between animal patients via contamination of the hospital environment, invasive procedures and veterinary personnel . As for ESBL-producing E. coli, it is important to limit the genetic pressure towards antimicrobial resistance development through microbiological monitoring of all cutaneous infections and rational use of antibiotics.
In respiratory diseases tetracyclines are the preferred choice, which is consistent with guidelines, with macrolides being second choice where penicillins should be. In reproductive infections penicillins and trimethoprim/sulphonamides are preferred by half of the respondents, as recommended by the guidelines; however, more than 83% of the practitioners stated a usual treatment duration of more than 8 days (vs. the recommended 5–7 days duration).
Regarding postoperative use of antibiotics in clean surgical procedures, half of the veterinarians reported their use in more than 50% of their cases, answering that they consider this to be “the typical procedure”. It is generally accepted that low-risk patients (healthy or with localised disease), and apyrexic patients with serious systemic illness, undergoing a clean or clean-contaminated procedure do not require antibiotic prophylaxis . Hence, it is a subject in which most of the respondents appear unaware of best practices. We must of course take into account that pet owners in Greece are usually surprised to learn that their pet will not receive antibiotics prior to routine surgery, and that they will be sent home without oral antibiotics.