The objective of this study is to survey general prescribing behaviour by Belgian companion animal veterinarians and to assess agreement of these practices with current treatment guidelines. Therefore an online survey was administered with five realistic and frequently occurring first-line cases to primary-care veterinary practitioners. For each case a predefined pattern of questions were asked about whether or not they would prescribe antimicrobials, if they would prescribe a non-antimicrobial treatment and if they would perform additional diagnostic steps. The responses were compared with recommendations in national guidelines and recent literature. The overall most prescribed antimicrobials were potentiated amoxicillin (43.0 per cent), fluoroquinolones (14.7 per cent), third-generation and fourth-generation cephalosporins (10.9 per cent) and tetracyclines (10.9 per cent). Only 48.3 per cent of the veterinarians complied with the guidelines in nearly all of the clinical scenarios (ie, prescribing antimicrobials when indicated, not prescribing antimicrobials when it is not indicated). Moreover, when prescribing highest priority critically important antimicrobials, susceptibility testing on bacterial cultures was performed in only 12.4 per cent of the prescriptions. The results showed that the prescribing behaviour of antimicrobial compounds by primary-care veterinary practitioners in dogs and cats is often not in agreement with national guidelines. Focus in improvement of this prescribing behaviour should be on performing the appropriate diagnostic steps and decreasing the use of highest priority critically important antimicrobials.
- antimicrobial guidelines
- companion animals
Statistics from Altmetric.com
Antimicrobial drugs are vital to animal and human health. However, antimicrobial use (AMU) is linked with the emergence of antimicrobial resistance (AMR).1–4 There is evidence that AMU in veterinary medicine is associated with AMR in human isolates.1 5–8 This poses a health and welfare risk for animals and human beings.9
In the last decade, numerous studies on AMU in production animals were performed.4 10–13 The data concerning AMU in dogs and cats remain scarce, despite the fact that antimicrobial drugs are the most commonly prescribed medicines in dogs and cats.14 15 The few available studies show that penicillins and broad-spectrum cephalosporins were most commonly used.15–19 These studies also indicated that, in spite of general recommendations,20 21 prescribing broad-spectrum potentiated amoxicillin was in most cases preferred over narrow-spectrum penicillins. Moreover, critically important antimicrobials,22 such as fluoroquinolones, were more frequently prescribed than antimicrobials with less importance for human health.16–18 23 When these critically important antimicrobials were prescribed, this was often without prior bacterial culture and susceptibility testing.15 16 18 A study about AMU in a veterinary teaching hospital reported that 38.4 per cent of the admitted dogs were treated with antimicrobials drugs without evidence of infection.24
In the last years, increasing efforts have been made to promote prudent prescribing behaviour and reduce antimicrobial consumption in veterinary medicine.25 This includes the development and distribution of AMU guidelines by national or private organisations in several countries.21 26–28 To be able to measure the effect of these strategies for antimicrobial stewardship, knowledge about current antimicrobial prescribing behaviour is essential.15 29
Therefore the aim of this study is to collect data on general antimicrobial prescribing behaviour in dogs and cats in Belgium and to assess to what extent these are in line with current therapeutic guidelines.
Materials and methods
A questionnaire was designed presenting five realistic and frequently occurring first-line cases. This questionnaire was presented to companion animal practitioners with the aim of assessing their clinical procedures and antimicrobial prescribing behaviour in these clinical scenarios.
In this cross-sectional study a survey was sent to 2363 veterinarians through the mailing list of the small animal clinic of the Faculty of Veterinary Medicine, Ghent University, whereof 1834 out of 2363 (77.7 per cent) veterinarians on the mailing list are Flemish (the Dutch-speaking part of Belgian). A detailed list of practising companion animal practitioners in Belgium is not available; overall there are 3469 veterinarians affiliated with the ‘Orde der dierenartsen’, which is a necessity to practise veterinary medicine. The online questionnaire was sent on January 6, 2016 with an accompanied letter to state the aim of the survey. The questionnaire could be filled in on a voluntary basis on SurveyMonkey (SurveyMonkey, San Mateo, California, USA) until the end of January, during which no reminders were sent.
The questionnaire was designed in collaboration with experts in companion animal internal medicine and based on two comparable studies regarding antimicrobial prescribing behaviour in dogs and cats.23 30 The questionnaire was pretested by five veterinarians and took approximately 15 minutes to complete. The full questionnaire is available as an online supplementary material.
Supplementary file 1
The first part of the questionnaire included a section with questions on the veterinarian, the veterinary practice and his/her general antimicrobial prescribing behaviour. This part consisted of eight closed and two semiclosed questions. The second part of the questionnaire consisted of five fictional clinical cases, three dogs (acute gastroenteritis, non-complicated urinary tract disease and acute tracheobronchitis) and two cats (upper respiratory tract disease (URTD) and skin wounds with systemic clinical signs). Details on these clinical scenarios are presented in Table 1. For each case, the signalment, history and general physical exam were presented and a predefined pattern of closed questions (‘Do you want to perform additional diagnostic steps?’, ‘Do you want to prescribe antimicrobials?’, ‘Do you want to prescribe a non-antimicrobial therapy?’) and semi-closed questions (‘What kind of additional diagnostic steps do you want to do?’, ‘What type of antimicrobial do you want to prescribe?’, ‘What type of non-antimicrobial therapy do you want to prescribe?’) were asked.
When asked about what type of antimicrobial treatment the respondent would prescribe, the respondent could select more than one antimicrobial drug and various forms of administration (eg, parenteral, oral). For instance, a respondent could choose to treat a dog with acute gastroenteritis with an injectable amoxicillin followed by oral amoxicillin. As such a combination of parenteral and oral antimicrobials or one of both could be selected.
Evaluation of use
The assessment of AMU was approached using the guidelines of the Center of Expertise on Antimicrobial Consumption and Resistance in Animals (AMCRA). In 2014 the AMCRA published national guidelines regarding AMU in dogs and cats. These are regularly updated and can be consulted online (www.e-formularium.be) by veterinarians free of charge after an online registration.21 Adherence to the guidelines is a free choice of the veterinarians: the recommendations are not legally binding. Per indication, information is provided about diagnostics, prevention, whether or not antimicrobials are indicated, and which antimicrobials are recommended for use (Table 1). When antimicrobials are indicated, they are colour-coded and listed as first, second or third choice. The colour codes of the antimicrobials (yellow, orange and red) represent the importance of the antimicrobial for public and animal health and are based on a merger of the lists produced by the World Health Organization (WHO)22 and World Organisation for Animal Health (OIE).31 Yellow antimicrobials have a narrow spectrum and are least important for public and animal health (eg, first-generation and second-generation cephalosporins, narrow-spectrum penicillins and sulphonamides). Red antimicrobials are most important for animal and public health and include third-generation and fourth-generation cephalosporins and fluoroquinolones, while orange antimicrobials are in-between and include, among others, tetracyclines, potentiated amoxicillin, broad-spectrum penicillins and doxycycline. Each colour code has also recommendations regarding the required accompanying diagnostic procedures.21 According to the AMCRA guidelines, it is recommended to perform additional diagnostic laboratory test(s) and antimicrobial susceptibility test(s) when using yellow antimicrobials. However, when using orange antimicrobials additional diagnostic laboratory test(s) should be done, and whenever red antimicrobials are used on top of these diagnostic laboratory test(s) antimicrobial susceptibility test(s) should be performed.21 Secondly, for each condition antimicrobials are listed as first-choice, second-choice or third-choice antimicrobials.
One of the outcomes of interest is ‘compliance with the guidelines’ and is defined as prescribing antimicrobials in the clinical scenarios where antimicrobials are likely indicated and refraining from prescribing antimicrobials when not indicated.
The results were exported to commercial software, Excel (Microsoft Corporation, Redmond, Washington USA). Initially the database was cleaned using the following procedures: (1) respondents who selected an antimicrobial treatment, but responded ‘no’ to the question ‘Do you want to prescribe antimicrobials’, had their answer changed to a ‘yes’; (2) respondents who selected a non-antimicrobial treatment, but responded ‘no’ to the question ‘Do you want to prescribe a non-antimicrobial therapy?’, had their answer changed to a ‘yes’; and (3) when the respondent did not select a fixed multiple choice answer, but answered in the ‘other section’ and this answer was similar to a fixed multiple choice answer, the latter was taken into account for further analysis.
After data cleaning, descriptive statistics were performed using commercial software, Excel and IBM SPSS statistics V.21.0 (SPSS, Armonk, New York, USA). Relative frequencies of AMU were quantified by counting the total of antimicrobial prescriptions per clinical scenario (denominator). For instance, in clinical scenario 1 fluoroquinolones for oral use were selected in 6 out of 15 prescriptions. Relative frequencies of diagnostic tests were quantified by counting the total of performed diagnostic tests per clinical scenario (denominator). The respondent could perform multiple diagnostic tests, so the total amount of ‘diagnostic tests performed’ could exceed the amount of respondents. The same reasoning was applied for ‘prescribed non-antimicrobial treatment’. Prevalence ratios (PR) and their corresponding 95 per cent confidence intervals (CI) were estimated to describe the relationship between compliance with the national guidelines or prescribing first-choice antimicrobials (dependent variables) and resources of information on the required therapies, such as scientific literature and permanent education (independent variables).
The response rate for the questionnaire was 12.0 per cent (n=284). Data on all five of the presented clinical scenarios were completed in 223 questionnaires (acute gastroenteritis n=264, non-complicated urinary tract disease n=243, acute tracheobronchitis n=236, URTD n=229 and skin wounds with systemic complaints n=223).
Information about the veterinary practice and non-case-specific antimicrobial prescribing behaviour
Most of the veterinarians worked in an exclusively companion animal practice (80.3 per cent). The veterinarians who worked in a mixed practice (19.7 per cent) predominantly treated dogs and cats (45.8 per cent of all treated species) or horses (29.7 per cent of all treated species). The majority of the veterinarians worked alone (52.1 per cent) and were female (60.6 per cent). There was a large variation in the years of work experience, with a mean of 19.4 (sd=12.2).
The majority of the veterinarians used continued education (96.8 per cent) and scientific literature (96.3 per cent) on a regular basis as resources of information on the required therapies. Nearly 74.7 per cent used information provided by representatives of pharmaceutical companies as a resource. The three most important factors in the decision-making process to prescribe antimicrobials were clinical presentation, owner compliance and ease of administration (99.3 per cent, 99.3 per cent and 95.9 per cent of the respondents, respectively, considered these as important). Financial restrictions and owner expectations (eg, the owner wants an injectable antimicrobial for his cat, instead of tablets) were reported as the least important factors (8.6 per cent and 14.2 per cent of the respondents, respectively, considered these as not important).
Treatment of acute gastroenteritis in a young adult dog
When presented the case about gastroenteritis, 64.7 per cent of the veterinarians would select additional diagnostic steps. Parasitic faecal examination was commonly selected (58.8 per cent of all performed diagnostic tests). Compliance with the guidelines and prescribed antimicrobials are presented in Table 2. For oral treatment, half of the prescriptions (10/20) were the extra-label use of the combination product metronidazole/spiramycin. Nearly all of the respondents (94.3 per cent) would prescribe a non-antimicrobial treatment, as stand-alone treatment or in addition to an antimicrobial treatment. There was a lot of variations between the types of non-antimicrobial treatment (eg, antiemetics, deworming, probiotics, adjusted diet).
Treatment of non-complicated urinary tract disease in a female neutered dog
The majority of the respondents (89.7 per cent) would select additional diagnostic steps. Microscopic examination of the urine (74.0 per cent of the veterinarians would select this), bacterial culture of the urine with susceptibility testing (51.5 per cent) and urine dipstick (49.4 per cent) were among the most selected diagnostic tests. Compliance with the guidelines and prescribed antimicrobials are presented in Table 2. Only 12.1 per cent of the prescriptions were first-choice antimicrobials, the majority (59.7 per cent) were prescriptions with second-choice antimicrobials and 27.2 per cent of the prescriptions were third-choice antimicrobials. Of the veterinarians who selected a non-antimicrobial treatment (65.0 per cent), the majority (81.0 per cent) chose NSAID.
Treatment of acute tracheobronchitis in a young adult dog
The minority of the veterinarians (25/240, 10.4 per cent) would select additional diagnostic steps when a dog with acute tracheobronchitis was presented, and 60 per cent of them (15/25) selected radiographic examination of the thorax. Compliance with the guidelines and prescribed antimicrobials are presented in Table 2. Whether treated with antimicrobials or not, a large part of the respondents (67.8 per cent) prescribed a non-antimicrobial treatment. NSAIDs (22.7 per cent), rest (21.5 per cent) and cough suppressant (20.9 per cent) were the top three non-antimicrobial treatments.
Treatment of acute URTD in a middle-aged cat
Only 14.2 per cent of the veterinarians would select additional diagnostic steps. Compliance with the guidelines and prescribed antimicrobials are presented in Table 3. Some prescriptions included topical ocular antimicrobial preparations containing gentamicin (9.5 per cent of the oral/topical prescriptions) and tetracycline (6.3 per cent of the oral/topical prescriptions). The majority of the respondents (67.3 per cent) would prescribe a non-antimicrobial treatment, such as NSAIDs (27.3 per cent of the non-antimicrobial treatments) or flushing of the eyes and/or nose with saline (22.6 per cent of the non-antimicrobial treatments). Thirteen out of 345 non-antimicrobial treatments (3.8 per cent) consisted of administering lysine, an amino acid.
Treatment of skin wounds in a 10-year-old cat with systemic complaints
When a clinical case regarding a cat with skin wounds with systemic complaints was presented, 85 out of 227 respondents (37.4 per cent) chose to select additional diagnostic steps, whereof needle puncture of the wound for cytological examination (75.5 per cent of the diagnostic testing) was by far the most selected. Only 15.3 per cent of the veterinarians would perform additional diagnostics, requested bacterial culture with susceptibility testing. Compliance with the guidelines and prescribed antimicrobials are presented in Table 3. Only 18.7 per cent of the prescriptions were first-choice antimicrobials, 44.9 per cent were second-choice and 34.3 per cent were third-choice antimicrobials. One-fifth (19.6 per cent) of the veterinarians who would prescribe an antimicrobial treatment would not select a supportive treatment. Regardless of any antimicrobial treatment, 79.4 per cent of the respondents prescribed a non-antimicrobial treatment. Irrigation and flushing of the wound (33.6 per cent of the non-antimicrobial treatments) and the administration of NSAIDs (29.3 per cent of the non-antimicrobial treatments) were the most performed. Some veterinarians drained and flushed the wound in combination with topical use of povidone-iodine (29 veterinarians, 16.4 per cent) or honey ointment (35 veterinarians, 19.8 per cent).
Overall prevalence and pattern of AMU
The overall most selected antimicrobials in dogs and cats in this survey were potentiated amoxicillin (43.0 per cent), fluoroquinolones (14.7 per cent), third-generation and fourth-generation cephalosporins (10.9 per cent) and tetracyclines (10.9 per cent) (Table 4). The majority of the prescriptions were orange colour-coded antimicrobials (68.2 per cent) and a quarter were red colour-coded antimicrobials (25.6 per cent). When red colour-coded antimicrobials were prescribed, bacterial culture with susceptibility testing was performed in only 12.4 per cent of the prescriptions. Only 18.6 per cent of the veterinarians complied with the guidelines in each answered clinical scenario and 48.3 per cent complied with the guidelines in four out of five cases. Veterinarians who used scientific literature as a resource for required therapies were more likely to prescribe according to the guidelines (PR=1.47, 95 per cent CI (1.03 to 2.11)). Furthermore, veterinarians who used continued education as a resource for required therapies were more likely to prescribe first-choice antimicrobials in the cases where antimicrobials were indicated (PR=1.84, 95 per cent CI (1.02 to 3.35)).
This survey provides data on AMU in dogs and cats in Belgium based on hypothetical clinical cases. By mailing the questionnaire we reached a substantial number of veterinarians in a short period of time and obtained numerous data on AMU and prescribing behaviour in dogs and cats. However, a low response rate was present, which could result in a selection bias. Besides, with this procedure, only a limited amount of information about the presented clinical scenarios could be given, which resulted in a simplification of the reality. Furthermore, social desirability response bias, the tendency of giving a ‘socially accepted’ answer, was likely to be present.32 Nonetheless, to our knowledge, this is one of the first studies in Belgium regarding AMU and prescribing behaviour in dogs and cats in Belgium. Although several campaigns were held to promote prudent AMU by providing national guidelines, it could be that not all of the respondents were familiar with these guidelines. Therefore when the respondents did not comply with the national guidelines, this could be due to ignorance or because the veterinarians deliberately deviated from the guidelines.
In the clinical scenarios in which an antimicrobial treatment was not indicated, many veterinarians would prescribe an antimicrobial treatment. However, between these clinical scenarios the percentage of veterinarians that would prescribe any antimicrobials varied greatly. Many veterinarians prescribed antimicrobials for a dog with acute tracheobronchitis or a cat with URTD, but for the clinical case about non-complicated acute gastroenteritis in dogs only 10.4 per cent of the respondents prescribed antimicrobials. This last percentage is substantially lower than in similar studies performed in Australia (59 per cent, 654 respondents) and the UK (25.5 per cent, 460 respondents), where a comparable case about non-complicated gastroenteritis was presented.23 30 This discrepancy between these studies could be because since the latter studies were performed, the awareness about prudent AMU increased. In the current study and in agreement with previous reports,15 30 antimicrobial drugs were frequently prescribed in case of acute tracheobronchitis in dogs (68.8 per cent) and URTD (43.7 per cent). Nonetheless, antimicrobial drugs are not indicated in acute uncomplicated tracheobronchitis in dogs, since mostly viral pathogens are the primary causal agent and since the disease is usually self-limiting.21 The results of this study suggest that for some conditions (ie, acute tracheobronchitis in dogs and URTD in cats) antimicrobial drugs are overused in dogs and cats, which is in agreement with findings in other studies.15 30
Scientific literature and continued education as resources for prescribing antimicrobial drugs were related to a higher compliance with the guidelines (ie, prescribing antimicrobials when necessary) and to prescribing first-choice antimicrobials more often instead of second-choice and third-choice antimicrobials. This was not unexpected, finding the results of a European survey study regarding antimicrobial prescribing behaviour where they reported that scientific literature and experience were the most important sources influencing the prescribing behaviour of the questioned veterinarians.25 This suggests that scientific literature and continued education are tools to influence the prescribing behaviour of veterinarians regarding antimicrobial drugs.25
Overall, in the current study, there was a tendency to prescribe antimicrobials without proper diagnostic work-up. Additional susceptibility testing is necessary for bacterial infections for which resistance is frequently reported. For example, bacterial culture with susceptibility testing should be performed in all cases of lower urinary tract infection in dogs.21 33 Despite the fact that nearly 90 per cent of the veterinarians performed additional diagnostic steps, only half of them performed bacterial culture with susceptibility testing of a urine sample as well. Secondly, when red colour-coded antimicrobials are prescribed, the majority of these prescriptions were not guided by bacterial culture and susceptibility testing. This seems to be consistent with other studies about antimicrobial prescribing behaviour in dogs and cats.15 16 18 29 34 A potential explanation could be that the owner refuses these tests because of the extra costs involved even though veterinarians did make that recommendation.18 More likely, veterinarians still underestimate the importance of prior bacterial culture and susceptibility testing when prescribing these red colour-coded antimicrobials, despite several campaigns to increase awareness. These two examples illustrate that there is still room for improvement regarding the diagnostic work-up in prescribing antimicrobials.
Compliance regarding supportive treatment was higher than compliance regarding diagnostic work-up. In the treatment approach of infected wounds, antimicrobials are only indicated in patients with systemic complaints (eg, fever)21 and complement local non-antimicrobial treatment when septicaemia is presumed. Many veterinarians would drain and flush the wound and some would apply topical povidone-iodine or honey ointment in combination with drainage and flushing of the wound. For this clinical scenario, most veterinarians would prescribe antimicrobials drugs in addition to a supportive treatment and not as stand-alone treatment. Further, also in accordance to the AMCRA guidelines,21 nearly all of the veterinarians would prescribe supportive treatment in dogs with acute gastroenteritis.
The present study revealed that the most prescribed antimicrobials for cats are potentiated amoxicillin and third-generation cephalosporins, whereas in dogs the most commonly prescribed antimicrobials are potentiated amoxicillin and fluoroquinolones. This is in agreement with a recent prospective study about antimicrobial consumption in primary-care veterinary practices in Belgium, where data on AMU were collected based on patient records.35 The latter study showed that in cats potentiated amoxicillin and cefovecin were the most prescribed antimicrobials, whereas in dogs predominantly potentiated amoxicillin and cephalexin were used. The observation that potentiated amoxicillin in dogs and cats16–19 23 and third-generation cephalosporins in cats17 19 are predominantly used is comparable with results from other countries. The frequent use of fluoroquinolones in dogs was also seen in a survey study in Italy16 and New Zealand.18
To minimise resistance selection both on pathogens and on commensal bacteria, veterinarians should select narrow-spectrum antimicrobials27 36 and therefore should select yellow colour-coded molecules when possible.21 The frequent use of broad-spectrum antimicrobials in this study is worrisome. It could be the result of uncertainty about the causal bacteria due to an incomplete diagnostic work-up by the veterinarians. However, also promotion by pharmaceutical companies might influence the common use of broad-spectrum antimicrobials.30 As a matter of fact, a substantial part of the respondents in this survey did rely on information given by representatives of pharmaceutical companies in their decision for prescribing antimicrobials. It has to be noted that not every use of broad-spectrum antimicrobials is misuse. For instance, doxycycline was frequently selected for respiratory conditions such as acute tracheobronchitis in a dog and feline URTD. In these conditions, this broad-spectrum antimicrobial may be an appropriate first choice. Bacterial agents associated with URTD in dogs and cats, such as Mycoplasma spp and Chlamydophila felis, are usually susceptible to doxycycline.37 38
Fluoroquinolones and third-generation and fourth-generation cephalosporins are listed as critically important drugs with highest priority by the WHO.22 Cefovecin, an injectable third-generation cephalosporin with prolonged action, was frequently used for the parenteral antimicrobial treatment in cats. This is likely linked to the ease of administration and the avoidance of poor owner compliance, which are considered as (very) important factors in the decision-making process of prescribing antimicrobials.30 Other researchers already raised their concerns about the wide use of highly critically important antimicrobials.15 16 19 39
In conclusion, the antimicrobial prescribing behaviour of Flemish companion animal veterinarians is often not in agreement with national recommendations. Focus should be on performing the proper diagnostic steps and decreasing the use of critically important antimicrobials, in particular those with highest priority. Further efforts have to be made to decrease the overall use of antimicrobials in companion animals and increase the awareness about justified AMU.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.