FEDERATION OF
infection societies
CONFERENCE 2018

13th – 15th November 2018 | Sage Gateshead – Newcastle

FEDERATION OF
infection societies
CONFERENCE 2018

13th – 15th November 2018 | Sage Gateshead – Newcastle

FEDERATION OF
infection societies
CONFERENCE 2018

13th – 15th November 2018 | Sage Gateshead – Newcastle

Posters

ANTIBIOTICS AND RESISTANCE ISSUES – Poster Nos. 026-067

026
The resistance mechanisms of multidrug-resistant Pseudomonas aeruginosa isolates to in vitro cetazidime/avibactam and ceftolozane/tazobactam, in Qatar

Abstract - 026

Poster 026

The resistance mechanisms of multidrug-resistant Pseudomonas aeruginosa isolates to in vitro cetazidime/avibactam and ceftolozane/tazobactam, in Qatar

Mazen A. Sid Ahmed1,2, Faisal A. Khan2, Hamad A. Hadi3, Muna Al-Maslamani3, Ali Sultan4, Bo Soderquist5, Emad Ibrahim1, Jana Jass2
1Microbiology Division, Laboratory Medicine and Pathology, Hamad Medical Corporation, Doha, Qatar. 2School of Sceince and Thechnology, Life Sceince, Orebro University, Sweden. 3Internal Medicine, Infectious Diseases, Hamad Medical Corporation, Doha, Qatar. 4Microbiology and Immunology, Weill Cornell Medicine-Qatar, Doha, Qatar. 5School of Medical Sciences, Faculty of Medicine and Health, Orebro University, Sweden

Introduction: Multidrug-resistant Pseudomonas aeruginosa (MDR-PA) infections are a serious clinical challenge due to their resistance to most available antibiotics and thus significantly limiting the treatment options. Recently, ceftazidime/avibactam (CZA) and ceftolozane/tazobactam (C/T) have been approved for clinical use and these combinations have demonstrated good activity against Gram-negative bacteria, including MDR-PA. However, resistance to these new antimicrobial combinations has already been reported. In this study, we aimed to characterize the genetic mechanisms driving resistance to CZA and C/T in MDR-PA.

Methods: From a total of 205 MDR-PA, 94 isolates were found to be resistant to one or both CZA and C/T antibiotics, of which 9 isolates (10%) were randomly selected for this study. The samples were isolated from different clinical specimen during 2014 and 2015. Antimicrobial susceptibility was determined by Etest. The 9 isolates were subjected to whole genome sequencing and bioinformatics analysis.

Results: The 9 isolates belonged to 5 different sequence types; ST-292 (n=1), ST-233 (n=1), ST-308 (n=3), and ST-823 (n=1) ST-2613 (n=3). The isolates had between 2 – 4 different b-lactamase genes from all classes. Four isolates had VEB-1a (Class A), 1 isolate had CARB-3, 4 isolates had VIM-2 (Class B), PDC-2, 3, 5, and 7 (Class C) were found in 3, 1, 1, and 4, respectively and OXA-4, 10, and 50 (class D) were found in 1, 2, and 9, respectively. The genes encoding the efflux pump regulators MexR, nalC, and nalD, as well as the efflux pump complexes, MexAB-OprM, MexCD-OprJ, MexPQ-OpmE and MuxABC-OpmB were found in all the isolates, however, the efflux pump regulator CpxR, soxR, and type B NfxB were detected in 7, 6, and 6 isolates.

Discussion: The presence of 2 or more different b-lactamase genes from all classes, including class B and D, which are known to be poorly inhibited by the b-lactamase inhibitors could explain the resistance of those isolates to CZA and C/T. In addition, the strains also contained four efflux pump complexes together with the two-component regulatory systems that control the expression of the efflux pumps. The relative contribution of the individual mechanisms in the various strains will need to be evaluated.

027
Evaluation of in vitro susceptibility of multidrug-resistant Pseudomonas aeruginosa isolates from Qatar to ceftazidime/avibactam and ceftolozane/tazobactam; β-lactam/β-lactamase inhibitor combinations

Abstract - 027

Poster 027

Evaluation of in vitro susceptibility of multidrug resistant Pseudomonas aeruginosa isolates from Qatar to ceftazidime/avibactam and ceftolozane/tazobactam; β-lactam/β-lactamase inhibitor combinations

Mazen A. Sid Ahmed1,2, Nahla Eltai3, Jamal Hamid1, Sulieman Abu Jarir4, Hamad A. Hadi4, Muna Al-Maslamani4, Hadi Yassine3, Ali Sultan5, Bo Soderquist6, Jana Jass2, Emad Ibrahim1
1Microbiology Division, Laboratory Medicine and Pathology, Hamad Medical Corporation, Doha, Qatar. 2School of Sceince and Thechnology, Life Sceince, Orebro University, Sweden. 3Bimedical Research Center, Qatar University, Doha, Qatar. 4Internal Medicine, Infectious Diseases, Hamad Medical Corporation, Doha, Qatar. 5Microbiology and Immunology, Weill Cornell Medicine-Qatar, Doha, Qatar. 6School of Medical Sciences, Faculty of Medicine and Health, Orebro University, Sweden

Introduction: The problem of multidrug-resistant Pseudomonas aeruginosa (MDR-PA) infections is a global healthcare challenge. Surveillance monitoring in Qatar established significant prevalence of MDR-PA. Collected isolates were testes against novel treatment options; ceftazidime/avibactam (CZA) and ceftolozane/tazobactam (C/T), which has been approved for treatment of complicated gram-negative infections.

Methods: A total of 205 MDR-PA isolates were collected between 2014-2015 from four hospitals in Qatar. The pathogens were isolated from: respiratory 44.9% (92), skin-soft tissues 26.3% (54), urine 23.4% (48), blood 3.4% (7) and other sites 2% (4). The activity spectrum of CZA and C/T were tested in vitro against MDR-PA using E-test according to standard recommendations.

Results: MDR-PA demonstrated favorable susceptibility to both CZA and C/T, at 68.8% (141/205) and 62.9% (129/205), respectively. Remarkably, 22.4% (46/205) of isolates were non-susceptible highlighting antimicrobial resistance endurance. Cumulative MIC distribution for CZA to both and C/T were MIC50/90, 4/64 µg/ml and MIC50/90, 2/256 µg/ml, respectively. When compared with eight other antibiotics, only colistin demonstrated higher susceptibility at 96.6%. The comparative results of phenotypically resistant of isolates to other antibiotics and CZA and C/T showed no significant correlation apart from fair agreement between C/T and CZA with amikacin (AMK) (0.37, p<0.001, k=0.27, p<0.001 respectively).

Discussion: The study demonstrated promising high in vitro susceptibility of CZA and C/T against MDR-PA isolates in Qatar. The results paved the way for potential future role of CZA and C/T in the management of MDR-PA infections and will be recommended as alternatives to complement existing options hindered by their recognized limitations.

028
Pre-admission antibiotic therapy and its effect on outcome in cases of mastoiditis: Implication for antibiotic stewardship

Abstract - 028

Poster 028

Pre-admission antibiotic therapy and its effect on outcome in cases of mastoiditis: Implication for antibiotic stewardship

Mitul Patel, Nikoleta Skalidi, Stephanie King, Konstance Tzifa
Birmingham Women’s and Children’s NHS Foundation Trust

Introduction: Mastoiditis is a complication of acute otitis media with potential for intracranial involvement and increased morbidity. There is concern that restricting antibiotic use for otitis media, as a result of improved antibiotic stewardship in primary care may lead to more complications of mastoiditis. The aim of this study is to investigate the effect of preadmission antibiotic therapy on prognosis in paediatric patients with acute mastoiditis and the differences in morbidity in patients less than two years olds.

Methods: Records from the Birmingham Children’s Hospital admissions database, from 2012-2016 inclusive, with diagnosis of acute mastoiditis were retrieved and reviewed retrospectively. The data were analysed for duration of symptoms, preadmission antibiotic therapy, delay in antibiotic administration, laboratory results, management of complications and duration of hospitalisation.

Results: 77 patients were included. 32 patients (41.5%) had received antibiotic treatment before admission while 45 patients (58.5%) had not. The mean duration of symptoms prior to the admission was 4.1days (SD=4.8) in the untreated group and 6.3 days (SD=4.7) in the group who received antibiotics. Preadmission antibiotic therapy did not significantly influence clinical outcomes (need for surgery or neurological complications) (p=0.915) or length of hospitalization (p=0.336) between the two groups. Similarly, preadmission antibiotics did not influence the clinical outcome in patients less than two years old (p=0.928). No significant difference was found in regards to complication rates in the age group of 0-23 months and 2-16 years (p= 0.657). Interestingly, patients who had a delayed prescription of antibiotics (after 4 days of symptoms) or did not receive a second line antibiotic due to treatment failure, were associated with complications (p=0.021) with a higher significance in the over two years old age group (p=0.003). High CRP values were observed in patients who had preadmission antibiotics but they were not predictive of a worse clinical outcome (p=0.665).

Discussion: Preadmission antibiotic therapy does not influence the clinical outcome in children with acute mastoiditis, including children <2 years of age. High CRP values do not predict a worse outcome. In those who received preadmission antibiotics, delay resulted in worse outcome. this patient group requires further investigation.

029
Double trouble

Abstract - 029

Poster 029

Double trouble

Celestine Eshiwe, Joshua York, Greta Johnson, Patrick Lillie
Hull and East Yorkshire Hospitals

Introduction: Antibiotic resistance is a major concern globally. Acinetobacter spp and Candida auris are organisms associated with significant antimicrobial resistance and have been associated with nosocomial outbreaks. They can be multidrug resistant and are highly transmissible between patients and the environment. Multi-drug resistant A. baumannii strains with additional resistance to carbapenems (MRAB-C) have been identified in several UK hospitals including paediatric settings. C. auris, since it was first identified in Japan, has continued to be a major concern; not just because it can cause fatal bloodstream infections that are difficult to treat, but also because it can be mis-identified in the laboratory causing a delay in initiating infection control practices which ensures that it can cause prolonged outbreaks. There is also little evidence that regular decontaminants like chlorhexidine are effective in killing the fungus. The screening process for both organisms is a little easier because similar sites (for instance, the nose, throat, perineum, wounds, sputa, tracheostomy, faeces, ante-cubital fossa) but the economic implications to the NHS cannot be ignored. To have both in one patient is certainly a management and infection control dilemma. This is not helped by the fact that there is no clearly defined approach for treating resistant organisms when there are few antibiotic options. It is often left to the discretion of the specialists.

Methods: A 56 year old man had road traffic accident in Nairobi, Kenya, and subsequently had craniotomy for brain haemorrhage/oedema, chest drains for haemothorax and pneumothorax and recieved antibiotics (Vancomycin and Meropenem) when he spiked temperatures while in hospital in Kenya.

He was stabilised and transferred to the UK and was isolated in an ICU room. C. auris was initially isolated from his blood culture and was only fully sensitive to amphotericin and flucytosine. Subsequently, A. baumannii was isolated which was resistant to all antibiotics tested including meropenem (tigecycline had the lowest MIC) except for colistin. Both organisms were also isolated from other body fluids (acinetobacter was identified in the pleural fluid, C. auris was also identified in the urine).

Two other patients acquired colonisation with acinetobacter which had the same genotype as that of our index case. Both were isolated and strict infection control practices were initiated. No secondary cases of C. auris were noted.

Discussion: High doses of caspofungin, and amphotericin treated the C. auris effectively. The inflammatory markers only improved after high dose meropenem was added and the colistin dose increased in addition to the high dose tigecycline. This illustrates the effectiveness of high dose combination therapy in situations of confirmed phenotypic resistance.

Strict infection control practices ensured that no new cases of C. auris were identified. The two patients that were colonised with acinetobacter were both isolated and staff practices, hand hygiene practices, use of fomites and equipment, and care bundles were reviewed and corrected. Screening and environmental sampling was negative and it was believed that staff spread was the likely reason for the spread. Terminal clean was done using hydrogen peroxide for the patient cubicles and subsequently the whole of the ICU.

Flucytosine, as well as change of the urinary catheter, was used to treat the urinary C. auris in the index case and topical terbinafine was used at cannula entry sites.

While MALDI-TOF correctly identified C. auris, The VITEK wrongly identified it as Candida haemulonii. The availability of the MALDI-TOF and its prompt use has ensured that the fungus was identified early to ensure that infection control acted swiftly to avert further spread of the organism.

030
Tedizolid: A service evaluation

Abstract - 030

Poster 030

Tedizolid: a service evaluation

Gavin Barlow1,2, Joshua York1
1Hull and East Yorkshire Hospitals NHS Trust. 2Hull York Medical School

Introduction: Tedizolid is a relatively new oxazolidinone antibiotic, licensed in the UK for acute skin and associated structure bacterial infections. Our centre has been using tedizolid to treat a wide range of infections, often switching from linezolid, and commonly using long courses from the outpatient antibiotic therapy service. This presentation demonstrates our experience.

Methods: We included all patients who received tedizolid between July 2016 and May 2017. We reviewed the documentation and system results, including diagnosis, length of treatment, and blood counts before and after, adverse reactions to treatment etc, together with basic patient parameters. We then collated the results and presented them.

Results: 17 patients were included, with prosthetic joint infections being the most frequent diagnosis (35%). 14 patients received linezolid for an average of 24 days, before tedizolid. Linezolid therapy resulted in an average drop of haemoglobin of 6 g/L with a drop in white cells of 1.9 x 109/L and a platelet drop of 149 x 109/L with 4 patients in the ‘critical’ range of having a haemoglobin <90. By contrast at the end of tedizolid therapy, haemoglobin had on average increased by 4 g/L, white cells had increased by 0.3 x 109/L and platelets had dropped by 15 x 109/L, with no patients finishing with a ‘critical’ haemoglobin <90 g/L. This was over an average of 32 days tedizolid therapy. 14 patients completed their planned course of tedizolid (82%), whilst 3 stopped early. 2 were stopped because of patient choice, whilst one was due to hyponatraemia. 13 patients (76%) had no adverse events related to tedizolid, whilst 1 had anaemia, 1 loose stools, 1 fatigue, and another had hyponatraemia. 11 of the patients showed improvement (64.5%), whilst 4 (23.5%) showed no change. 2 (12%) patients worsened whilst on tedizolid and 1 of these died.

Discussion: From our results, tedizolid appears to be tolerated well; only 3 patients stopped treatment early, and only 4 patients had adverse events; of these; the patient with anaemia received a treatment break for 40 days, during which time his haemoglobin recovered; he then tolerated a long course of tedizolid. The patient complaining of loose stools seems a genuine adverse event, whilst the other patient who stopped her tedizolid by choice due to complaints of fatigue had recently stopped taking her analgesia. The patient with the documented hyponatraemia has tolerated subsequent courses of tedizolid.

Patients were treated for a range of complicated, polymicrobial infections. Most patients showed an improvement whilst on tedizolid (65%). Although many of these required ongoing antibiotics as part of their continuing treatment; this would certainly suggest efficacy.

2 patients (12%) worsened on the tedizolid; one of these had cerebral aspergillosis, in which tedizolid was given for a possible superadded bacterial infection. He sadly deteriorated and died of the aspergillosis. The other patient was the same mentioned above with the discitis; she felt much worse and was recorded as a worsening; but this should be considered along with her cessation of analgesia.

Our data shows that those who do not tolerate linezolid can switch to tedizolid and tolerate a long course of this. This builds on earlier literature showing that switching from linezolid to tedizolid in linezolid related myelotoxicity results in improved blood counts.1

Overall, our evaluation would support the use of tedizolid for long term use in a wide range of indications, particularly bone and joint, and its usefulness in switching from linezolid.

1 Khatchatourian L, Le Bourgeois A, Asseray C et al. Correction of myelotoxicity after switch of linezolid to tedizolid for prolonged treatments. Journal of Antimicrobial Chemotherapy. 2017 April; 72(7): 2135-2136

031
Vancomycin quality improvement: introducing a prescription chart across an NHS England Trust

Abstract - 031

Poster 031

Vancomycin quality improvement: introducing a prescription chart across an NHS England Trust

Dominic Waugh1, Laura Cunliffe1, Andy Karvot2, Josh York3
1Diana Princess of Wales Hospital, Grimsby. 2North Lincolnshire & Goole NHS Trust. 3Hull Royal Infirmary

Introduction: Vancomycin is a drug with a narrow therapeutic index, associated with dose limiting ototoxicity and nephrotoxicity. In North Lincolnshire and Goole (NLAG) NHS Foundation Trust, vancomycin dosing is calculated according to guidelines outlined within “Path Links Antibiotic Formulary and Prescribing Advice for Adults”. While the advice detailed for vancomycin prescribing in this policy is comprehensive, it is currently prescribed on a standard hospital drug chart. This practice is contrary to Path Links partner United Lincolnshire Hospitals (which currently have a dedicated vancomycin prescription chart). A previous audit at NLAG had been undertaken to establish whether there was benefit to introducing a dedicated vancomycin prescription chart. Of the patients suitable in this audit for inclusion (N=14), only 57.1% had received a loading dose, of which only 75% were correct. Initial maintenance dose was correct in only 15.4% of cases. The main recommendation from this project was to introduce a dedicated vancomycin treatment sheet.

Aims:

  1. Design and introduce a standard vancomycin prescription chart across NLAG NHS Trust.

    2. Ensure full compliance with NLAG ‘Medicines & Therapeutics Committee’. 

    3. Seek feedback from junior doctors across NLAG regarding vancomycin charting and prescribing.
    4. Educate doctors, nurses and pharmacists prior to introduction of a dedicated vancomycin prescription chart across NLAG.

Plan:

  1. Sample NLAG junior doctors thoughts on vancomycin prescribing. 

    2. Design a vancomycin prescription chart for use across NLAG. 

    3. Achieve approval for use of chart through Medicine & Therapeutics Committee with Chief Pharmacist. 

    4. Print charts and distribute across NLAG. 

    5. Discuss introduction with NLAG staff.

Methods: A questionnaire regarding vancomycin prescribing in NLAG was distributed to junior doctors. Trainees (N=12) were asked to rate their confidence in vancomycin prescribing using a Likert scale 1 – 5 (not confident – very confident). Junior doctors in NLAG were surveyed before and after introduction of a standard vancomycin prescription chart with education regarding it’s use. 

Results: Responses indicated trainees were generally not confident in prescribing vancomycin in NLAG. It was generally agreed that introduction of a standard vancomycin prescription chart across the Trust would make prescriptions safer. Trainees (N=15) were asked to rate their confidence in vancomycin prescribing after evaluation of the new vancomycin chart using the same scale. Using unpaired T test, pre and post chart responses showed statistically significant (P=<0.0001) increased average confidence levels (mean pre-chart confidence – post chart confidence = – 2.00, 95% confidence interval of difference: from -2.82 to -1.18). Trainees showed a statistically significant increase in confidence levels when prescribing vancomycin using the new prescription chart.

Discussion: Introduction of a standardised prescription chart across a trust can be feasibly achieved over a 3-4 month time frame. The process took longer than anticipated due to a number of factors. This includes obtaining approval from appropriate individuals in: medicines & therapeutics, key pharmacist roles, design and reprographics, governance and trust documentation standards. Engaging junior doctors in design and implementation of a vancomycin prescription chart can demonstrate a significant improvement in confidence when prescribing vancomycin and should increase patient safety. Trainees commented that they felt the information they needed to prescribe vancomycin safely and confidently was now easier to interpret with the introduction of the new prescription chart. Introduction of a dedicated antibiotic prescription chart has the potential to make prescribing of less frequently encountered antibiotics safer for patients and contribute to good antimicrobial guardianship and practice. It is imperative that pharmacy ensures prescriptions for vancomycin are dispensed only when charted correctly. Due to the relatively low number of prescriptions of vancomycin across NLAG annually, this should be re-audited in around 12 months times to evaluate whether there has been significant improvement in prescribing of initial / loading dose of vancomycin.

032
Outcomes of point-of-care influenza and respiratory syncytial virus testing in an emergency department

Abstract - 032

Poster 032

Outcomes of point-of-care influenza and respiratory syncytial virus testing in an emergency department

Stuart Bond, Jade Lee-Milner, Christine Cruise, Sarah Robertshaw, Neil Ullyott, Emma Godfrey
Pinderfields Hospital, Wakefield

Introduction: Respiratory viruses account for a large proportion of acute respiratory admissions in adults.1 Traditional laboratory testing resulted in local delays of 48 hours as samples were processed at a neighbouring hospital. Increased diagnostic speed while maintaining testing accuracy may lead to improved patient outcomes as well as reduced inpatient admissions, ward closures, antibacterial requirements and hospital costs.

Aim: To evaluate the effectiveness of a point-of-care testing (POCT) model for influenza A/B and respiratory syncytial virus (RSV) in an NHS Trust without onsite laboratory-based virus diagnostics.

Methods: A retrospective before-and-after intervention study was conducted to compare outcomes between off-site and POCT for influenza/RSV. A polymerase chain reaction (PCR) influenza A/B and RSV POCT machine (Cepheid®) was introduced to the emergency department (ED) of Pinderfields Hospital in November 2017. POCT was available for all inpatients, and full respiratory virus screens were processed off-site if requested. ED nurses underwent training to ensure consistent nasopharyngeal sampling and use of POCT equipment. Patients were selected for testing based on likelihood of requiring treatment according to Public Health England guidance. Study patients were identified from pathology records for the time periods 1 December 2016 to 28 February 2017 and 1 December 2017 to 28 February 2018. Positive influenza A/B and RSV tests were compared for: length of hospital stay (LOS), length of ED stay, hospital costs, antibacterial duration, ratio of oseltamivir use to positive tests, and inpatient mortality. Hospital outbreaks and test costs were also evaluated. Ethics approval was not required. A chi-square test was used for proportions and Mann-Whitney U-test for LOS comparison.

Results: A total of 1557 tests (349 for 2016/17; 1208 for 2017/18) were identified. Results for positive influenza A/B tests where all data were available were analysed following removal of duplicates (n= 32) and RSV tests due to low numbers (n=5 in 2016/17; n=34 in 2017/18). Median LOS decreased from 8.5 days (IQR 4-21 days; n=46) in 2016/17 to 4 days (IQR 2-8; n=346; p=0.12) in 2017/18. Median length of ED stay was also reduced by 42 minutes (6h 42m [IQR 3h 54m – 10h 15m] to 6h 0m [4h 0m – 9h 0m]). Median hospital deficit reduced from -£1,310 (IQR -£3,394 – £632) in 2016/17 to -£381 (-£1,145 – £395) in 2017/18. Median antibacterial duration reduced from 6 days (IQR 2.5-7) to 4 days (IQR 2-6). The ratio of oseltamivir use (full courses) to positive tests reduced from 386/46 (8.4) in 2016/17 to 758/346 (2.1) in 2017/18. There were three inpatient influenza outbreaks in 2016/17 and none in 2017/18. Test costs were equivalent with both methods at £36 each. Inpatient mortality was 15% (7/46) in 2016/17 and 7% (26/353) in 2017/18 (p=0.07).

Discussion: Introduction of a POCT method for influenza was associated with improvements in patient-related outcomes. Reduced time to results allowed for more targeted oseltamivir use, and use of single inpatient rooms for influenza positive patients resulted in a reduction in hospital outbreaks. Limitations included: retrospective before-and-after design, improved outcomes potentially related to improved test availability meaning more patients with mild disease were tested, lack of availability of clinical information due to paper notes and lack of linkage of clinical systems, and low patient numbers in the first year reducing statistical power. Further research is warranted using prospective methodology to investigate the practicality and cost of different respiratory virus testing methods.

  1. Brendish NJ, Malachira AK and Clark TW (2017). Molecular point-of-care testing for respiratory viruses versus routine clinical care in adults with acute respiratory illness presenting to secondary care: a pragmatic randomised controlled trial protocol (ResPOC). BMC Infectious Diseases. 17:128.

033
Antimicrobial prescribing behaviour in the Emergency Department

Abstract - 033

Poster 033

Antimicrobial prescribing behaviour in the Emergency Department

Kathryn. M Ashton, Dorothy. J Frizelle, Stuart.E Bond, Sarah Robertshaw
Mid Yorkshire Hospitals NHS Trust, Wakefield

Introduction: The prevalence of antimicrobial resistance is rising with over-use and misuse of antimicrobials continuing to be a key driver for the emergence of resistance. The Emergency Departments (EDs) at Mid Yorkshire Hospitals have been identified as high antimicrobial users. As the interface between inpatient and community settings the behaviour of prescribers in the ED has significant impact on antimicrobial prescribing across the healthcare economy.

Uncertainties about pathogenesis, heavy workflow and patient’s desire for antimicrobial therapy have been cited as motives for increased antimicrobial prescribing in existing research and while prescribing behaviours have been explored in the literature there are limited studies that specifically examine antimicrobial prescribing behaviours in the unique environment of EDs.

The aim of this study was to identify and understand the behavioural factors that influence antimicrobial prescribing in our own ED in order to enable the introduction of successful initiatives to reduce antimicrobial use.

Methods: A short survey was compiled with the contribution of a clinical psychologist. A literature review was undertaken to identify comparable studies and assist survey design. The survey consisted of agreement with 13 statements, assessed using a 4-point Likert scale (‘strongly agree’ to ‘strongly disagree’). Statements were based on the principles of The Theory of Reasoned Action to explore beliefs, motivation and attitude towards antimicrobial prescribing. Information on length of practice and grade was included. The survey was issued to prescribers at Pinderfields ED by an antimicrobial pharmacist. Survey participants did not receive prior information about the survey and all responses were anonymous. Responses were collated by our data analyst.

Results: There were 14 consultant, 6 registrar, 15 junior doctor and 3 non-medical prescriber responses collected, 38 overall. Mean duration of practice was 8.6 years, 2 participants omitted years in practice. 98% (n=37) of participants agreed that they consider AMR when prescribing antimicrobials, with 100% agreeing appropriate antimicrobial prescribing is important to them. 100% of participants disagreed they prescribe antimicrobials due to patient expectation. 11% (n=4) agreed with the statement that they prescribe antimicrobials when they do not have time to counsel the patient. 30% (n=11) agreed where they are unsure of a diagnosis it would be preferable to prescribe an antimicrobial, while 82% (n=31) agreed the risk of the patient deteriorating influenced their decision to prescribe an antimicrobial. 87% (n=33) agreed senior colleagues would support their prescribing decisions. 98% (n=37) agreed feedback on their prescribing would be useful to them.

Discussion: Overall the responses from the survey suggest that whilst prescribers in the ED believe that appropriate prescribing of antimicrobials is important it is temporally separated from moment-by-moment decisions in clinical situations. The major barrier to appropriate antimicrobial prescribing was identified as the risk of a patient deteriorating and uncertainty around diagnosis. This desire to avoid risk can be linked to the theory of loss aversion and suggest that these prescribers are behaving in a natural way.

This study did have limitations; the small number of participants meant that is was not possible to identify significant differences between the responses of the different grades of prescribers or duration of experience. Also responses were not statistically validated.

Results of the survey highlighted enablement strategies are key in the ED to improve antimicrobial prescribing. Following dissemination of survey responses, timely feedback and post prescription review have been identified by ED clinicians as a preferred intervention to improve confidence and positively re-inforce a culture of appropriate antimicrobial prescribing among the team. Although the majority of responders said they felt supported in their decision making, it is key that senior prescribers encourage and educate junior colleagues to enable them to reduce the perceived risk of not prescribing antimicrobials.

034
Analysis of re-prescribing of antibiotics used for urinary tract infection in the community using routinely collected national information

Abstract - 034

Poster 034

Analysis of re-prescribing of antibiotics used for urinary tract infection in the community using routinely collected national information

William Malcolm1, Alistair Beith2, Jacqueline Sneddon3
1Health Protection Scotland, NHS National Services Scotland, Glasgow. 2Information Services Division, NHS National Services Scotland, Glasgow. 3Scottish Antimicrobial Prescribing Group, Healthcare Improvement Scotland, Glasgow

Introduction: Lower urinary tract infection (UTI) is common in women, and a common reason for empirical community antibiotic use. A 3-day course of trimethoprim or nitrofurantoin for non-pregnant women of any age with acute lower UTI is recommended. Trimethoprim is used first line in most NHS boards in Scotland. In 2016, trimethoprim non-susceptibility in E. coli urine isolates was stable though high at 36.7%, with 97.3% susceptible to nitrofurantoin. Samples submitted are biased towards non-susceptibility only being submitted in complicated cases or on failure of empirical treatment.

We aimed to:

  • identify the proportion of women with a prescription for nitrofurantoin or trimethoprim who received a further antibiotic prescription within 7 days of initial course completion.
  • determine if women who received 3-day courses had more repeat courses than those with 5 or 7-day courses.

Methods: Records from Prescribing Information System (PIS) E-messaging (national data on medicines dispensed in the community) were linked to SMR-01 data (national hospital activity data) to:

  • identify females aged ≥16years with a community prescription for nitrofurantoin 50mg tablets/capsules or 100mg modified release capsules or trimethoprim 200mg tablets from 1 January to 31 December 2016
  • exclude patients who had antibiotics prescribed or a hospital admission in the previous 90 days.

The outcome of interest was the proportion of individuals who received a further prescription for a pre-specified range of antibiotics commonly used for UTI within 7 days of the date of initial course completion.

Logistic regression tested the effect of course length on repeat prescription. Age group; Charlson score (prior 5 years); number of medicines by legacy BNF paragraphs prescribed in prior 12 months; care home status and exposure to antibiotics (DDDs) in the prior 12 months were used as additional predictors.

Results: Out of 233,248 individuals, 144,004 (61.8%) met the inclusion criteria: 35,387 (24.6%) were prescribed nitrofurantoin and 108,617 (75.4%) trimethoprim. The number of patients included was similar across all age groups. Only 1% of patients were care home residents. Around half of patients had an unknown Charlson score and 40% had a score of zero.

In females who received trimethoprim, 92.8% did not have a further prescription for UTI antibiotics within 7 days of initial course completion. In those receiving nitrofurantoin, 94.0% had no further UTI antibiotics prescribed within 7 days.

After adjusting for other predictors, 3-day courses of trimethoprim were not associated with an increased rate (7.0%, p=0.15) of repeat prescribing compared to 5 or 7-day courses (7.5%, 8.1%). Seven day courses were associated with a statistically significant increase in repeat antibiotics compared to 3-day courses (p<0.001).

In those who received nitrofurantoin after adjusting for other predictors 3-day courses were associated with an increased rate (6.5%, p<0.01) of re-prescribing compared to 5 or 7-day courses (5.3%, 5.2%).

Discussion: This study using routine national data indicates that despite the reported high levels of trimethoprim non-susceptibility in E. coli urine isolates only 7.2% of adult females returned within 7 days of the date of completion of the course for a further prescription for a UTI antibiotic. Moreover only 6.0% of females who received nitrofurantoin had a further UTI antibiotic within 7 days.

These data should provide reassurance to clinicians that trimethoprim and nitrofurantoin remain appropriate first choice empirical antibiotics in females aged ≥16years with symptoms of simple uncomplicated lower urinary tract infection.

Three day courses of nitrofurantoin but not trimethoprim were associated with higher rates of repeat antibiotic prescribing compared to 5 or 7-day courses. However the absolute difference was low.

035
The rise of pharmacist prescribing of antibiotics in Scotland and implications for antimicrobial stewardship in Scotland

Abstract - 035

Poster 035

The rise of pharmacist prescribing of antibiotics in Scotland and implications for antimicrobial stewardship in Scotland

William Malcolm1, Rita Nogueira2, Jacqueline Sneddon3
1Health Protection Scotland, NHS National Services Scotland, Glasgow. 2Information Services Division, NHS National Services Scotland, Edinburgh. 3Scottish Antimicrobial Prescribing Group, Healthcare Improvement Scotland, Glasgow

Introduction: The antimicrobial stewardship programme coordinated by the Scottish Antimicrobial Prescribing Group (SAPG) supports optimisation of antibiotic use through reducing inappropriate and unnecessary antibiotic use. Antibiotic use in primary care has been reducing since 2012. In 2016 the rate of antibiotic use was lower than at any point since 2004. Prescribing has mainly been undertaken by GPs but over the past 5 years the number of non-medical prescribers, particularly nurses, has increased year on year.

Pharmacy First, a Scottish Government scheme, aims to improve patient access to GP appointments by encouraging patients with certain minor ailments to use their community pharmacy for treatment rather than making a GP appointment. All pharmacists involved undertake education and training to support their new role. Following some pilot work the scheme was rolled out across Scotland in 2017 focused on management of uncomplicated urinary tract infection (UTI) in women aged 16 to 65 years and impetigo in any age group. Community pharmacists carry out a patient consultation in the pharmacy and provide advice and treatment if required under locally agreed patient group directions (PGD). If antibiotics are supplied they are prescribed on NHS pharmacist prescription forms which feed into the national datamart on medicines use in Scotland. The service is available both within GP opening hours and out of hours.

We investigated recent trends in antibiotic prescribing by pharmacists.

Methods: Data on dispensed antibiotic prescriptions written by pharmacists on NHS Community Pharmacy prescription forms (CPUS, CP1 and CP2) from 1st January 2013 to 31st December 2017 were obtained from the Prescribing Information System, a database of all NHS prescriptions dispensed in Scotland held by NHS National Services Scotland.

Results: In 2017 there were 16,031 antibiotic prescriptions written by community pharmacists dispensed in Scotland. Although this represents only 0.4% of antibiotic use in primary care, pharmacist prescribing of antibiotics has increased greatly since 2013 as shown in table 1.

 

Year

Number of prescriptions

2013

1,476

2014

1,867

2015

3,000

2016

8,070

2017

16,031

 

In 2017, 82% of pharmacist antibiotic prescriptions were for trimethoprim. This mainly reflects the roll out across Scotland of the Pharmacy First UTI consultation service. Some NHS boards have also established local schemes where patients can consult community pharmacist for: skin infections; exacerbation of chronic obstructive pulmonary disease (COPD); and for Chlamydia testing and treatment.

The other frequently prescribed antibiotics included: flucloxacillin (3.2%), the antibiotic of choice for skin infections; amoxicillin (3.0%), first line for COPD; and azithromycin (2.8%), first line for Chlamydia.

Discussion: The National Clinical Strategy for Scotland (http://www.gov.scot/Publications/2016/02/8699) has highlighted that primary care will be at the centre of transforming service delivery in NHS Scotland. As part of this transformation it is likely that community pharmacists using their clinical skills will increasingly work as part of multidisciplinary teams in primary care to optimise medicines use. The Pharmacy First scheme is a first step towards all pharmacists being trained as independent prescribers.

It is important that this new group of prescribers is adequately supported and included in a team approach to antimicrobial stewardship. SAPG is working with the Royal Pharmaceutical Society Scottish Branch to develop and deliver interactive education sessions to increase pharmacists’ confidence around providing advice for patients about self-care of infections and ensuring appropriate use of antibiotics.

036
Risk factors for resistant bacteraemia in patients with suspected sepsis: cohort study using linkage of routinely collected national data

Abstract - 036

Poster 036

Risk factors for resistant bacteraemia in patients with suspected sepsis: cohort study using linkage of routinely collected national data

William Malcolm1, Eilidh Fletcher2, Julie Wilson1, Chris Robertson1,3, Charis Marwick4
1Health Protection Scotland, NHS National Services Scotland, Glasgow. 2Information Services Division, NHS National Services Scotland, Edinburgh. 3University of Strathclyde, Glasgow. 4Population Health & Genomics, School of Medicine, University of Dundee

Introduction: Essential to initial treatment of suspected sepsis is early administration of appropriate antibiotics. Initial antibiotic treatment is usually empirical and choice based on local prescribing guidelines. These guidelines generally do not take account of patient characteristics and the likelihood of antibiotic resistance in individual patients. The empiric regimen for initial treatment of sepsis on admission to hospital varies amongst NHS boards in Scotland but the most commonly recommended combination is amoxicillin plus gentamicin plus metronidazole. We aimed to characterise risk factors for having a bacteraemia non-susceptible to this regimen among patients who have a blood culture taken on admission to hospital.

Methods: Blood culture and susceptibility data were obtained from the Health Protection Scotland Electronic Communication of Surveillance in Scotland (ECOSS) dataset. Patient-level hospitalisation data were obtained from National Services Scotland (NSS) General/Acute Inpatient dataset (SMR01) and community NHS prescriptions from NSS Prescribing Information System (PIS). In Scotland all individuals have a patient identifier, the Community Health Index (CHI) number, which enables record linkage across multiple datasets. Public Benefit and Privacy Panel approval to link data was obtained.

We identified patients aged ≥16 years with a blood culture result in ECOSS from 01/01/2016 to 30/09/2017. We defined suspected sepsis as having had a blood culture taken, irrespective of result. Using culture and susceptibility results, positive cultures were defined as non-susceptible if the isolate was resistant to both amoxicillin and gentamicin, or resistant to metronidazole. Susceptible organisms and negative blood cultures were combined to form the comparator group.

Where more than one sample was taken during a hospital admission, de-duplication based on the European Centre for Disease Control protocol for the submission of data to the European Antimicrobial Resistance Surveillance Network was conducted. Only one sample per admission, the ‘most resistant’ sample, was included in the analysis.

Blood culture data were linked using CHI to SMR01 and PIS to determine demographics, co-morbidity and previous antibiotic exposure. Samples taken from day three onwards were excluded. Multivariable logistic regression was undertaken to determine risk factors for non-susceptibility.

Results: A total of 102,760 samples were analysed: 1,064 (1.0%) were non-susceptible to the empiric regimen. Of those, 839 (79%) were non-susceptible to amoxicillin and gentamicin and 225 (21%) to metronidazole.

Multivariable logistic regression found increasing age associated with non-susceptibility. Those aged ≥85 years were 3.57 (95%CI 2.15-6.39) times more likely to be non-susceptible compared to those aged 16-24, as were males (OR 1.24; 95%CI 1.10-1.40) and those residing in a care home (OR 2.01; 95%CI 1.66-2.43). Increasing comorbidity was associated with increased likelihood of non-susceptibility – those with a Charlson score of ≥5 were 1.44 (95%CI 1.15-1.79) times more likely to be non-susceptible than those with a score of zero and those with ≥4 previous hospitalisations were 2.61 (95%CI 2.11-3.22) times more likely than those with no previous hospitalisations. 

Exposure to any antibiotic in the community in the previous three months was found to be associated with non-susceptibility and those previously prescribed an antibiotic were 1.20 times (95%CI: 1.06-1.37) more likely to be non-susceptible. 

Discussion: Risk factors for non-susceptibility to the commonly used antibiotic regimen for sepsis have been characterised. Exposure to any antibiotic in the three months prior to admission is a risk factor for non-susceptibility; next we will examine cumulative exposure to all antibiotics and to specific antibiotics in the three, six and 12 months prior to blood culture. Following this, we aim to develop risk prediction models as the next step towards developing a clinical decision support tool to help clinicians identify patients at risk of being non-susceptible to the commonly recommended empirical treatment for sepsis in NHS Scotland.

037
Impact of intravenous antimicrobial dose reductions on clinical outcomes in patients presenting with sepsis-induced acute kidney injury (AKI)

Abstract - 037

Poster 037

Impact of intravenous antimicrobial dose reductions on clinical outcomes in patients presenting with sepsis-induced acute kidney injury (AKI)

Sabha Mahmood1, Stephen Hughes2
1Imperial College London. 2Chelsea and Westminster NHS Foundation Trust, London

Introduction: Sepsis-induced acute kidney injury (AKI) requires early, effective antimicrobial therapy to reduce mortality. Pathophysiological changes in sepsis lead to altered antimicrobial pharmacokinetics. Sepsis increases capillary permeability, resulting in increased volume of distribution of hydrophilic antimicrobials and the possibility of subtherapeutic concentrations. Treatment of sepsis-induced AKI requires a balance between treatment failure and toxicity. No validated guidelines exist regarding antimicrobial dosing in this state. Current guidelines regarding antimicrobial dosing adjustments are obtained from pharmacokinetic studies in patient populations with normal kidney function or chronic kidney disease (CKD). These populations have stable kidney function and do not reflect the acute changes seen in AKI patients. We aimed to investigate the effect of antimicrobial dose reductions on clinical outcomes in patients with sepsis-induced AKI. Secondary objectives were to determine the impact of development of AKI and severity of AKI on outcomes in septic patients.

Methods: Retrospective single-centre cohort study conducted at Chelsea and Westminster Hospital including patients >18 years with microbiological confirmation of sepsis, defined as blood culture indicating Gram-negative bacteraemia, between December 2015 to March 2018. Institutional databases were used to collect data on patient demographics, comorbidities, admission/discharge dates, transfer to ITU, date of death and prescription and dose of antimicrobials. Data collected on kidney function included serum creatinine values and estimated glomerular filtration rate (eGFR). Acute rise in serum creatinine was used to identify and stage patients with AKI according to RIFLE classification. There were 109 patient cases of AKI. Collected data was compared between groups based on development of AKI, severity of AKI and effect of antimicrobial dose reductions.

Results: Patients who received antimicrobial dose reductions did not have significant differences in clinical outcomes compared to patients receiving full dose. However, a non-significant trend towards increased length of stay by 10.5 days and increased mortality (33% compared to 10% with full dose) was observed in patients who received dose reductions. Median eGFR at time of blood culture was lower in the group who received a dose reduction (23 ml/min/1.73m2) compared to the group who received full dose (39 ml/min/1.73m2) (p=0.014). Median serum creatinine value for patients receiving full dose antimicrobial was 169 µmol/L at time of AKI and 70 µmol/L 6-10 days after positive blood culture (p<10-4).

The incidence of AKI in patients with Gram-negative sepsis was 24.1% in 2017 and development of AKI was associated with increased length of stay by 5.5 days (p=0.001). Greater severity of AKI was associated with older age by 5.5 years (p=0.032) and increased length of stay by 6.5 days (p=0.019).

Discussion: Although no significant difference between patients with Gram-negative sepsis and AKI receiving full antimicrobial dose or reduced dose was observed, a trend for worse clinical outcomes was seen with antimicrobial dose reductions. Inability to achieve statistical significance may be due to a low number of patients receiving dose reductions. This is reflected by the presence of local guidelines at the study institution recommending full dose antimicrobials for patients with sepsis-induced AKI. Patients with a lower eGFR were more likely to receive a reduced antimicrobial dose. Use of eGFR values to guide dosage adjustments has been validated in CKD patients but do not accurately reflect acute kidney dysfunction. Administration of full-dose antimicrobial did not worsen kidney function and no cases of beta-lactam induced seizures were reported. Development of AKI and more severe AKI led to longer length of hospital stay. The results of this study will need to be replicated in a prospective multicentre study.

038
A national survey to determine current practice regarding antimicrobial dosing in patients with sepsis-induced acute kidney injury (AKI)

Abstract - 038

Poster 038

A national survey to determine current practice regarding antimicrobial dosing in patients with sepsis-induced acute kidney injury (AKI)

Sabha Mahmood1, Stephen Hughes2
1Imperial College London. 2Chelsea and Westminster NHS Foundation Trust, London

Introduction: There are no validated guidelines for dosing of antimicrobials in patients with sepsis-induced acute kidney injury (AKI). Pathophysiological changes in sepsis lead to altered antimicrobial pharmacokinetics. Sepsis increases capillary permeability, resulting in increased volume of distribution of hydrophilic antimicrobials. Inadequate antimicrobial dosing in this state contributes to poorer patient outcome, with underdosing leading to treatment failure as well as contributing to antimicrobial resistance. Current guidelines regarding antimicrobial dosing are obtained from pharmacokinetic studies in patient populations with normal kidney function or chronic kidney disease (CKD). These patient populations have stable kidney function and hence do not reflect the acute changes in kidney function occurring in AKI patients. We aimed to survey current practice regarding antimicrobial dosing in the United Kingdom and determine the resources commonly used to guide dosing.

Methods: A questionnaire was developed via the online SurveyMonkey® platform and sent to the infection, renal and critical care networks of the United Kingdom Clinical Pharmacy Association (UKCPA) in March 2018 to determine whether trusts had local guidelines for antimicrobial dosing in patients with varying kidney function. A clinical vignette of a septic patient with AKI was used to assess pharmacists’ recommendations for dosing of intravenous co-amoxiclav, meropenem, gentamicin and ciprofloxacin.

Results: A total of 73 pharmacists from 51 separate trusts responded. The respondents were from the following departments: intensive care (42%), microbiology/laboratory/infectious diseases (29%), renal medicine (18%), acute or specialist medicine (7%), surgery (1%) and other (3%). 93%, 50% and 23% of respondents had local trust guidelines for antimicrobial dosing in patients with normal kidney function, CKD and AKI, respectively. Of the respondents with local antimicrobial dosing guidelines for AKI patients, 50% were from intensive care, 44% were from microbiology/laboratory/infectious diseases and 6% were from renal medicine. Based on the clinical vignette, 57%, 63%, 32% and 21% of respondents recommended a reduced antimicrobial dose of co-amoxiclav, meropenem, gentamicin and ciprofloxacin, respectively, before 48-hour review. 41% of respondents recommended avoiding gentamicin and 1% recommended avoiding ciprofloxacin. The remainder recommended prescription of full dose antimicrobial (according to Summary of Product Characteristics (SPC) from product licencing information). Resources used by respondents to guide antimicrobial dosing included Renal Drug Handbook (88%), SPC (74%), local guidelines (66%), expert opinion (63%), BNF (37%) and primary evidence (26%).

Discussion: Less than a quarter of respondents are part of trusts with local guidelines for antimicrobial dosing in patients with AKI. The majority of respondents with local trust guidelines for antimicrobial dosing in AKI patients were from the intensive care department or microbiology/laboratory/infectious diseases departments. There is wide variability in current practice regarding antimicrobial dosing in septic patients with AKI. Respondents were more likely to recommend a reduced dose of co-amoxiclav, meropenem and gentamicin compared to full dose. This variability is reflected by the differences in resources used to guide antimicrobial dosing. Resources commonly used by respondents to guide antimicrobial dose adjustments in renal impairment are based on creatinine clearance or estimated glomerular filtration rate (eGFR) values. Although validated in CKD patient populations, these values do not accurately reflect the extent of acute kidney dysfunction. There is a need for national guidelines to standardise antimicrobial dosing in critically ill patients with a view to optimise dosing to prevent antimicrobial resistance and treatment failure.

039
A qualitative survey of Antimicrobial Stewardship Practice across secondary care in London

Abstract - 039

Poster 039

A qualitative survey of Antimicrobial Stewardship Practice across secondary care in London

Stephen Hughes1, Preet Panesar2
1Chelsea & Westminster NHS Hospitals, London. 2University College London Hospital

Introduction: Antimicrobial stewardship (AMS) focuses on optimisation of antimicrobial prescribing to improve patient outcomes, whilst minimising the development of antimicrobial resistance (AMR) for our patients and wider society. An emergent AMS strategy has developed rapidly in response to national incentives, historically focusing upon C. difficile infection (CDI) rates in secondary care and more recently in through national priorities incentivising a reduction in antimicrobial usage in primary and secondary care. This rapid development has resulted in a huge variation in AMS activities with minimal overlap in services as a result. To map current AMS service across London and identify the AMS service provisions provided a mixed methods survey was conducted among infection specialist pharmacists

Methods: An electronic survey was constructed to derive a measure of frequency with which different AMS interventions were being undertaken. Survey items were selected based upon existing literature, including the Cochrane review on AMS interventions. The 12-question closed response survey was circulated in February 2018via the Chief Pharmacist group to all acute NHS Trust providers in London requesting local service feedback from an appropriate AMS team member.

Results: Twenty-three acute NHS trust replied (100% return rate) and all surveys were fully completed (100% completion rate) and were included for analysis.

AMS strategies for restricting antimicrobial availability were commonly encountered for 1st or subsequent antimicrobial dosing in 65% of Trusts. Most services (70%) had ability to monitor ‘protected’ antimicrobial following supply/prescription to review appropriateness using dispensing records and/or electronic prescribing systems. Daily AMS activities, such as multi-disciplinary ward-round reviews of adult wards (91%), paediatric wards (70%) and bacteraemia review rounds (42%), are the norm. Antifungal rounds are evident in 35% of Trusts at least weekly. The scope of all these wards rounds ranges from targeted wards to hospital-wide practice.

Novel diagnostic markers used to guide antimicrobial therapy were infrequently used in practice. Procalcitonin use to guide initiation (9%) and duration (30%) of therapy was not common. Antifungal antigens were more frequently utilised (52%).

Where electronic prescribing (EP) was available to an AMS team (16 trusts), it was used in 88% of these organisations for retrospective patient surveillance and in 38% for real-time patient review.

Penicillin allergy referral services on behalf of the AMS service were only available for some or all patient groups in 34% of trusts.

Clinical governance reporting duties of the AMS services vary across Trusts in terms of the breadth of reporting and frequency. Auditing of documented indication of treatment (47%), 48-72hour review outcome (52%) and time to 1st dose (74%) was completed at least quarterly in the majority. Auditing total treatment duration (51%), drug therapeutic monitoring (78%), surgical prophylaxis (96%) and empirical treatment compliance (70%) was completed less frequently, often dependent on staffing availability.

Outcome measures for AMS interventions were typically completed at Trust-wide level for raw antimicrobial usage [DDD/1,000admissions] (90%) and CDI rates (96%). Prescriber level reporting of these outcomes (13%), Trust level resistance trends (70%), and patient related outcomes measures (mortality, length of stay, re-admissions) (8%), were infrequently reviewed.

Collaboration between AMS services across care boundaries was evident but wide variations were reported. Specialist pharmacist contribution to primary-care guidelines was common [83%]; activity related to mapping local resistance trends (44%), education and training (9%), and attendance to local primary care AMS meetings (30%) was variably reported.

Discussion: Extensive variation in AMS practices among acute Trust is seen across London. The emergent strategies of local teams to firstly act against the CDI burden and more recent aligning to national priorities, including CQUIN targets, results in an extensive variation in AMS practices of acute Trusts even in small geographical area like London

040
An evaluation of hyperkalaemia and acute kidney injury associated with In-hospital co-trimoxazole dosing

Abstract - 040

Poster 040

An evaluation of hyperkalaemia and acute kidney injury associated with in-hospital co-trimoxazole dosing

Stephen Hughes, Francis Edozie, Nabeela Mughal
Chelsea & Westminster NHS Hospitals, London

Introduction: Moves to avoid broad-spectrum beta-lactam prescribing has seen an upsurge in co-trimoxazole usage. Ongoing concerns with co-trimoxazole associated hyperkalaemia and kidney injury exist. The impact of dosing on this remains unclear and high-risk groups are not well defined in the literature. To describe the dosing and patient-risk factors associated with co-trimoxazole treatment complications a retrospective observational study was undertaken in a single London Teaching hospital.

Methods: All adult-patients receiving co-trimoxazole for 4 or more days therapy at a dose 960mg/day or more between February 2016 and April 2018 were identified from electronic prescribing records. Patients receiving treatment as an out-patient or on the intensive care unit were excluded. Electronic health records were used to extract relevant pathology results and concurrent medications. Changes in serum potassium from baseline to peak during treatment was calculated, as was estimated glomerular filtration rate (eGFR) using the Modification of Diet in Renal Disease equation, enabling identification of Acute kidney injury (AKI) using RIFLE classification following change in eGFR from baseline. Linear regression was undertaken to identify variables associated with elevated serum potassium and AKI using GraphPad PRISM.

Results: 208 patients were identified with a mean age of 62 years (range 20-98 years) and 59% were male. 159 patients had follow-up serum potassium levels available, of which 40(25%) had a potassium rise of 1mmol/L or greater whilst on treatment (max rise of 2.5mmol/l to peak of 6.4mmol/L).

Variables independently associated with increased serum potassium included age >60 years (odds ratio [OR] 4.83; 95% CI 2.2-10.64; P=.0001), concurrent acute kidney injury (AKI) (OR 3.79; 95% CI 1.42-10.56; P= .008), and doses ≥1.92g/day co-trimoxazole prescribed (OR 5.13; 95% CI 2.1-12.5; P= .0003). Dosing of 960mg BD, 1.44g BD and 60-120mg/kg/day all increased potassium level compared to low-dose <1.92g/day (OR 5.31; 95% CI 2.0-14.1; P= .0008, OR 4.9; 95% CI 1.28 – 18.86; P= .02 and OR 4.9; 95% CI 1.64-14.77; P= .045 respectively). Concurrent ACEi/ARB (OR 1.07; not-significant (ns)), aldosterone antagonists (OR 0.52; ns), presence of chronic kidney disease (CKD CrCl<60ml/min) (OR 1.0, ns) and treatment duration did not impact serum potassium levels.

Follow up kidney function was available for 146 patients in this study. A total of 14.3% of this study group had a documented acute kidney injury (AKI) in the first 10 days of treatment, classified as risk (9.5%) or injury (4.8%) in line with RIFLE criteria. Variables independently associated with AKI included concomitant prescribing of an ACEi/ARB (OR 10.37; 95% CI 3.24-33.25; P= .0001), age >60 years (OR 5.13; 95% CI 1.77-14.92; P= .003) and co-trimoxazole dosing ≥1.92g/day. Dosing of 1.92g/day, 2.88g/day and 60-120mg/kg/day all increased the incidence of AKI compared to doses <1.92g/day (OR 6.16; 95% CI 1.58-23.96; P= .009, OR 6.1; 95% CI 1.08 – 34.56; P= .04 and OR 4.84; 95% CI 1.06-22.02; P= .04 respectively). Baseline CrCl <60ml/min (CKD 3a – 5) (OR 2.28; 95% CI 0.77-6.70; P= .134), concurrent aldosterone antagonists, and length of therapy did not significantly impact on AKI risk with co-trimoxazole treatment. Limitations exist as no other patient co-morbidities could be identified in this study.

Discussion: Kidney function and serum potassium should be monitored following initiation of co-trimoxazole treatment. An association between co-trimoxazole and AKI appears more common among older patients, those concurrently prescribed ACEi/ARB therapy, and co-trimoxazole doses of ≥1.92g/day. A rise in serum potassium can be predicted at higher doses of co-trimoxazole and among elderly patients, but not in those who are concomitantly on ACEi/ARB agents, or those with CKD.

041
Comparison of Mast D72C AmpC ESBL and carbapenemase detection set with the D68C AmpC and ESBL, D63C ESBL, and D69C AmpC detection sets for the identification of ESBL and AmpC activity in Enterobacterales species

Abstract - 041

Poster 041

Comparison of Mast D72C AmpC ESBL and carbapenemase detection set with the D68C AmpC and ESBL, D63C ESBL, and D69C AmpC detection sets for the identification of ESBL and AmpC activity in Enterobacterales species

Fiona Shaw, Teresa Barton, Pamela Hulme, Jane Johnson, Norma Thompson, James Cargill
Alder Hey Children’s Hospital, Liverpool

Introduction: Alder Hey is a specialist paediatric hospital in Liverpool, providing clinical services across north-west England and north Wales. Basic microbiology and virology services are provided by the on-site laboratory.

Clinical and surveillance isolates of Enterobacterales species are screened for suspected ESBL and AmpC production as below:

  • Isolates resistant to cefpodoxime, cefoxitin, cefotaxime or ceftazidime by EUCAST disc diffusion susceptibility testing
  • Isolates resistant to cefotaxime or ceftazidime by Vitek AST card susceptibility testing
  • Isolates inherently resistant to cefoxitin (chromosomal AmpC carriers)
  • Isolates are screened with either the Mast D68C AmpC and ESBL detection set (suspected E. coli isolates) or both the D63C ESBL and D69C AmpC detection sets.

The D72C AmpC ESBL and carbapenemase detection set was evaluated as a possible replacement for all three tests in use.

Methods: Isolates from samples submitted for the surveillance of multi-resistant Gram-negative carriage (rectal swab, faecal samples, and throat swabs) were processed as normal; those isolates which underwent testing for suspected ESBL or AmpC expression were tested concurrently with the D72C disc set.

Results: 82 isolates had D72C and D63C/D68C/D69C results for comparison.

AmpC and ESBL negative: 7 isolates (3 E. coli, 3 K. pneumoniae, 1 E. asburiae) tested negative for AmpC and ESBL expression by D72C; these results were confirmed by D68C or the combination of D63C and D69C tests.

AmpC and ESBL positive: 6 isolates (3 E. cloacae complex, 1 each of C. freundii, K. pneumoniae and Pantoea spp.) tested positive for AmpC and ESBL expression by D72C; these results were confirmed by the combination of D63C and D69C tests.

AmpC positive: 33 isolates (15 E. cloacae complex, 10 E. coli, 6 C. freundii, 1 Serratia spp., 1 Pantoea spp.) tested positive for AmpC expression by D72C; these results were confirmed by D68C or D69C tests.

Inducible AmpC: 7 isolates (6 E. cloacae complex, 1 C. freundii) were reported with inducible AmpC by D72C; AmpC expression was reported by D68 or D69C tests.

ESBL positive: 24 isolates (14 E. coli, 7 K. pneumoniae, 2 K. oxytoca) tested positive for ESBL expression by D72C; 23 results were confirmed by D68C or D63C with the exception of 1 K. oxytoca isolate (negative by both D63C and D69C tests).

Suspected Carbapenemase: 2 isolates (both E. cloacae complex) were reported with suspected carbapenemase activity by D72C; both isolates showed AmpC expression by D69C and were repeat isolates of known KPC-expressing E. cloacae complex organisms.

Equivocal: 3 isolates (1 each of C. freundii, K. oxytoca, and M. morganii) were reported equivocal by D72C; AmpC expression was identified for the C. freundii and M. morganii isolates by D69C tests.

Overall: 76 of 82 isolates (92.7%) showed agreement between the different test, with 5 isolates (6.1%) requiring additional work following the D72C test and 1 (1.2%) showing a discrepancy between the tests.

Discussion: The D72C AmpC ESBL and carbapenemase detection test showed equivalent detection as the D63C/D68C/D69C tests for the detection of AmpC and ESBL activity, with the advantage that the single test can identify inducible AmpC alongside derepressed and acquired enzymes, unlike the D68C.

We decided that the small number of isolates reported as “Equivocal” or “Suspected Carbapenemase” by D72C should be followed up with D63C and D69C tests, with those showing suspected carbapenemase activity also receiving CPE screening.

042
Community screening to determine the prevalence of carbapenemase producing organisms in East London, 2018

Abstract - 042

Poster 042

Community screening to determine the prevalence of carbapenemase producing organisms in East London, 2018

Jennifer Henderson1,2, Holly Ciesielczuk1, Shona Nelson2
1Barts Health Trust, London. 2University of the West of England, Bristol

Introduction: Carbapenems are broad spectrum antibiotics reserved for patients who are extremely ill or suspected of having an infection caused by a multidrug resistant organism. Over the past ten years there has been a dramatic increase in resistance to carbapenems, seen worldwide, which is a growing cause for concern.

Carbapenemases are enzymes, produced by an array of common Gram-negative organisms, which hydrolyse this class of antibiotic, conferring resistance. The main protagonists are the “Big five” carbapenemases, KPC, OXA-48, IMP, VIM and NDM, which have been reported across the UK. However, these reports are often a result of reactive screening, outbreaks, inpatient surveillance and from diagnostic samples. To date there have been no studies investigating the prevalence of carbapenemase-producing organisms (CPO) in the UK community.

Methods: This study was performed at Barts Health NHS Trust (BHT), the largest trust in the UK, which serves 2.5 million patients across three London boroughs: Tower Hamlets, Newham and Waltham Forest. A total of 200 non-duplicate community stool samples, received sequentially by the Microbiology Laboratory at BHT, were included in this study. Patient age, sex and foreign travel history were extracted from the laboratory information management system (LIMS), enabling the identification of potential risk factors for CPO carriage.

Screening was performed by transferring a pea-sized portion of stool into nutrient broth and enriching overnight at 37°C. The broth was sub-cultured onto mSuperCARBA selective medium and incubated for a further 18 – 24 hours at 37°C.

Colonies were identified by MALDI-TOF. All identified Enterobacteriaceae, Acinetobacter species and Pseudomonas species underwent antibiotic susceptibility testing (AST) by disk diffusion, according to EUCAST guidelines, against meropenem, ertapenem, fosfomycin, mecillinam, amikacin, temocillin and piperacillin-tazobactam. All isolates, regardless of the AST results, were tested for possession of carbapenemase genes using the published in-house RT-PCR assay.

Results: Of the 200 community samples tested, only 1 patient tested positive for a CPO (NDM-producing Pseudomonas aeruginosa). Of note, this patient had travelled to the Caribbean. Of the 199 who screened negative for possession of a CPO, 46 also had foreign travel listed, with the most common countries being Turkey, Morocco and Bangladesh. A total of 122 had no travel history detailed and 31 had no clinical details provided. Patients ranged from 1-93 years of age.

Discussion: PHE guidelines state that patients from a high-risk geographical locations such as Bangladesh, India, South East Asia, Italy, Turkey, Greece and Israel are at risk of CPO carriage and infection. At BHT, a significant proportion of our patient population originate from these high-risk locations and 22/46 of our study patients visited them in the last 12 months. However, only one CPO was detected in our community, giving a prevalence of just 0.5%. Furthermore, the CPO was detected in a patient who had travelled to the Caribbean, suggesting that we need to reconsider who is high-risk for CPO carriage and the relevance of national CPO rates, at least at a local level. In addition to foreign travel, previous hospitalisation is also considered a risk factor; however this cannot be determined from this study.

Given the low CPO detection rate in this study, it could be expanded to include a larger sample size, which would enable the confirmation of the community prevalence rate observed here.

The stool samples used in this study were from patients presenting with a suspected gastrointestinal infection. This could lead to an imbalance of bacteria and therefore not be truly representative of the normal gut microbiome of that individual. To overcome this, future work could also include otherwise healthy patients for a more representative sample of the community.

043
Evaluating the use of the RPS antibiotic checklist by community pharmacists for potential impact on AMR

Abstract - 043

Poster 043

Evaluating the use of the RPS antibiotic checklist by community pharmacists for potential impact on AMR

Atiqah Liaqat1, Hava Nasar1, Mariam Qasim1, Aida Shiraz1, Gill Hawksworth1, Saima Afzal1, Sarah Frank1, Philip Howard2
1University of Huddersfield. 2NHS Improvement Patient Safety Team, London

Introduction: Antimicrobial resistance (AMR) is a worldwide public health crisis. This study analyses how community pharmacies could deliver elements of the 5-year AMR strategy as set out by the Department of Health in 2013¹ by using the Royal Pharmaceutical Society (RPS) antibiotic checklist² when counselling patients. Evaluation of the use of the checklist and counselling on AMR following its adoption will identify if any changes are required.

Objectives: To evaluate the use of the RPS antibiotic checklist by community pharmacists for counselling when dispensing antibiotic prescriptions in conjunction with the appropriate Public Health England “treat your infection” patient information leaflet (PIL). Views of community pharmacists were assessed using a questionnaire.

Methods: This study required and received ethics approval. A pilot antibiotic counselling checklist developed in 2017 was revised and promoted by RPS in their 2017 national antimicrobial stewardship campaign. Pharmacists who had used the pilot version were re-trained by researchers. The RPS checklist was used for 4 weeks in January 2018 to counsel patients when antibiotics were dispensed in conjunction with a PHE PIL.. A self completion tally sheet recorded the number and individual counselling points used with these patients. A piloted questionnaire recorded usefulness of checklist and PIL.

Results: Twelve pharmacists had consented to use the RPS checklist and tally sheet but after 4 weeks only 7 had recorded any data. A total of 211 patients were counselled, 29.8% (n=63) on alcohol, 31.2% (n=66) on not sharing with family and 42.6% (n=90) were asked to return unwanted medicines. A further 73.4% (n=155) were counselled on side-effects, 92.4% (n=195) on how to take correctly, 82.9% (n=175) on course length and 81.5% (n=172) on finishing the course. Almost all (95.7% n=202) were asked the reason for the antibiotic, and 85.3% (n=180) were told the likely duration of the infection. A PHE leaflet was given to 28.4% (n=60) of patients, and 45.9% (n=97) were given safety-netting advice (seeking further advice) and 29.3% (n=62) on self-care. Only 18.4% (n=39) were advised on flu vaccination. The same number told about hand hygiene despite it being a key element of the RPS national campaign about antimicrobial resistance. From the questionnaire, 71% (n=5) of pharmacists routinely used the RPS checklist. 57% (n=4) found it quite useful and 43% (n=3) thought it time consuming. Barriers included ‘too many options’, ’very busy in the pharmacy’, ’we know the standard advice to give on antibiotics’ and ‘all pharmacy bags state –don’t share medication’ as other reasons. The low number of eligible pharmacists taking part was a limitation.

Discussion / Conclusion: This small study suggests community pharmacists are delivering some recommended educational elements about AMR to patients but not all. A review of the RPS checklist to make it less time consuming might encourage more pharmacists to use it in practice when counselling on antibiotic use. With more prominence given to AMR and specifically hand hygiene, it could help deliver elements of the Government’s antimicrobial resistance strategy¹.

References

¹ Department of Health. 2013. UK 5 Year Antimicrobial Resistance Strategy. https://www.gov.uk/government/publications/uk-5-year-antimicrobial-resistance-strategy-2013-to-2018 (accessed 27 May 2018)

² Royal Pharmaceutical Society Antimicrobial stewardship quick reference guide :RPS website https://www.rpharms.com (accessed 27 May 2018)

044
Evaluating the impact of the KAW antimicrobial campaign among the public

Abstract - 044

Poster 044

Evaluating the impact of the KAW antimicrobial campaign among the public

Atiqah Liaqat1, Hava Nasar1, Mariam Qasim1, Aida Shiraz1, Gill Hawksworth1, Saima Afzal1, Sarah Frank1, Philip Howard2
1University of Huddersfield. 2NHS Improvement Patient Safety Team, London

Introduction / Background: Antimicrobial resistance (AMR) is a worldwide public health crisis. As part of the 5-year AMR strategy set out by the Department of Health in 2013 ¹, Public Health England launched a campaign designed to raise awareness on appropriate antibiotic use. This study analyses how the eight- week national campaign known as the ‘Keep Antibiotic Working’ (KAW) campaign from 23rd of October 2017 (Public Health England 2017²) impacted on the knowledge of the public.

Objectives: To measure the knowledge of the public visiting community pharmacies following the KAW campaign.

Methods: This study received ethics approval. The data was obtained by approaching the public in busy local areas. A questionnaire was completed face to face aimed at both male and female population over 18years. KAW had targeted all adults especially focusing on women aged 20-45, men and women aged 50+. Data was collected January to February 2018. The study purpose was explained, and a subject information leaflet given plus some AMR information

KAW promotional material, leaflet and poster were shown to interviewees.

Results: Of the 102 questionnaires completed, 46% were male (72% West Yorkshire and 28% Sheffield/Manchester) and 54% were female (67% West Yorkshire and 33% Sheffield /Manchester). When asked about where to seek advice on antibiotics/infections 42% (56/123) from GP, 30% (37/123) would self-care but only 5% (6/123) would see a community pharmacist. Further options were NHS111, Dentist and Walk-in centre and less than 4% used the internet.

For advice on antibiotics 77% (79/102) would never ask a community pharmacist, and 10% stated once a year or when required. 92% of interviewees correctly identified that fungal and viral infections did not need antibiotics and 60% correctly identified bacterial infections can be treated with antibiotics.

17% West Yorkshire and 35% Sheffield/Manchester were aware of the national KAW campaign. Manchester previously had the same pilot campaign the year before.

15% of responses were from the targeted 25-35 year age group, 13% from 36-46 years age groups and 37% of total responses being from 18-24 age group (so not population representative). These are all potential parental age groups. Of the 27% (10/37) in 18-24 age group who were aware of the campaign, 70% saw it on TV, 60% posters, 30% billboards and 10% via social media. The TV was the most common source by 36-46 age group at 50% (3/6), but posters were for 75% (3/4) of 47-56 age group.

Before the KAW campaign, only 7% knew a lot, 43% knew something about AMR, 20% had only heard of it and 30% has never heard of it.

Asked if the campaign had changed their thoughts on antibiotics, 63% said yes. In addition when asked how they would change their behaviour or use of antibiotics 22% (54/245) would now always finish a prescribed course and an additional 22% said they would seek a pharmacist’s advice on antibiotics compared to 28% who would see a doctor. This is encouraging as promotional material for the KAW campaign did not mention pharmacists.

Discussion / Conclusion: The KAW campaign demonstrated an improvement in the public’s understanding about antibiotics. It was mainly seen on TV or posters. The study suggests that following the KAW campaign there may be a change in behaviour of the public around finishing the course of antibiotics and seeking advice from pharmacists which may have a future impact on AMR. Repeat exposure to a campaign appears to increase recall.

Reference

¹ Department of Health. 2013. UK 5 Year Antimicrobial Resistance Strategy. https://www.gov.uk/government/publications/uk-5-year-antimicrobial-resistance-strategy-2013-to-2018 (accessed 24 May 2018)

² Public Heath England (n.d). Keep Antibiotics Working . Retrieved from https://campaignresources.phe.gov.uk/resources/campaigns/58-keep-antibiotics-working/Overview (accessed 24 May 2018)

045
Analysis of the use of educational material available for community pharmacists from the KAW campaign

Abstract - 045

Poster 045

Analysis of the use of educational material available for community pharmacists from the KAW campaign

Atiqah Liaqat1, Hava Nasar1, Mariam Qasim1, Aida Shiraz1, Gill Hawksworth1, Saima Afzal1, Sarah Frank1, Philip Howard2
1University of Huddersfield. 2NHS Improvement Patient Safety Team, London

Introduction: Antimicrobial resistance (AMR) is a worldwide public health crisis. As part of the 5-year AMR strategy set out by the Department of Health in 2013 ¹, Public Health England (PHE) launched a campaign designed to raise awareness on appropriate antibiotic use. This study assesses how the educational materials available during the eight- week national ‘Keep Antibiotic Working’ (KAW) campaign from 23rd of October 2017 (Public Health England 2017²), were used and impacted on the work of community pharmacists.

Objectives: To obtain feedback from the community pharmacy teams regarding the educational materials related to antimicrobials in the KAW antimicrobial campaign.

Methods: This study received ethical approval. Piloted questionnaires were posted to community pharmacies throughout West Yorkshire at the beginning of January 2018 for a 4 week data collection period. A pre-paid envelope and subject information sheet about the KAW campaign and project was included and replies were monitored. Following an initial low number of responses, pharmacies who had not responded were re-sent the questionnaire and information material.

Results: A total of 107 questionnaires were posted out and 24 (22%) were returned. When asked if the team was aware of the 8 week KAW campaign in October 2017, 92% said yes, and 58% saw the campaign saw it on TV. The campaign was also seen at the Chemist and Druggist website and via promotional material sent by the local council and by NHS England. 83% (20/24) of pharmacies participated in the campaign.41% of the pharmacies stocked the Keep Antibiotic Working leaflets, 66% displayed KAW campaign posters. 29% only spoke to patients about the campaign. However 8% used their own leaflets, as it was difficult to get campaign material from the PHE website. Of those that did not take part, 8% did not know about it, and 8% did not order any materials. There was some engagement from patients: 25% of pharmacies said patients had asked about the campaign. The community pharmacists did not think that the KAW campaign had increased patient interest about antibiotic use: 38% thought it had not improved and 50% indicated they thought it had not made any difference. Overall, only 13% of pharmacists thought the campaign was worthwhile.

Discussion / Conclusion: From a small sample of community pharmacies, most had taken part in the KAW campaign. A small proportion had not heard of the campaign or had not registered and to get materials from the PHE website to run the campaign.

 

References

¹ Department of Health. 2013. UK 5 Year Antimicrobial Resistance Strategy. https://www.gov.uk/government/publications/uk-5-year-antimicrobial-resistance-strategy-2013-to-2018 (accessed 24 May 2018)

² Public Heath England (n.d). Keep Antibiotics Working . Retrieved from https://campaignresources.phe.gov.uk/resources/campaigns/58-keep-antibiotics-working/Overview (accessed 24 May 2018)

046
BSAC Bacteraemia Resistance Surveillance: A five year review of non-susceptibility (2013-2017)

Abstract - 046

Poster 046

BSAC Bacteraemia Resistance Surveillance: A five year review of non-susceptibility (2013-2017)

Carolyne Horner1, Shazad Mushtaq2, BSAC Standing Committee on Resistance Surveillance1
1British Society for Antimicrobial Chemotherapy, Birmingham. 2Public Health England, London

Introduction: The British Society for Antimicrobial Chemotherapy (BSAC) Bacteraemia Resistance Surveillance Programme has monitored the antimicrobial susceptibility in the major organisms causing clinically significant bacteraemia in the UK and Ireland since 2001. This update reviews the latest five years of data (2013-2017).

Methods: Consecutive isolates causing clinically significant bacteraemia throughout the UK and Ireland were tested, with participating laboratories (n=24-40) collecting 7-20 isolates/species per year. Minimum inhibitory concentrations were determined centrally by BSAC agar dilution.

Results: 16,198 isolates were tested comprising: Staphylococcus aureus (n=2476); coagulase-negative staphylococci (CNS) (n=1094); Enterococcus spp. (n=1230); Streptococcus pneumoniae (n=1154); β-haemolytic streptococci (n=1221); α-/non-haemolytic streptococci (n=1026); Escherichia coli (n=2611); Klebsiella spp. (n=1264); Enterobacter spp. (n=1012); Proteeae (n=1153); Serratia spp. (n=791), and Pseudomonas spp. (n=1166).

Among S. aureus, the MRSA rate decreased by 50% (12%, 2013 to 6%, 2017, p=0.003) and high-level mupirocin resistance, mediated by mupA, was rare (n=3 MRSA, n=6 MSSA). Thirteen species of CNS were represented: most isolates were S. epidermidis (65%, n=706). Rates of methicillin resistance and mupirocin resistance in CNS were 71% and 18%, respectively, and did not change significantly over the five year period; however, non-susceptibility to teicoplanin increased (15% in 2013 to 24% in 2017, p=0.01).

There were 48 serotypes of S. pneumoniae: three isolates were non-typeable. The most common serotypes were 8 (16%), 12F (9%), 22F (8%), 2 (7%) and 19A (6%). Non-susceptibility to penicillin in S. pneumoniae was 5.6%: the majority (15/53) were serotype 15A. Among other α-/non-haemolytic streptococci (representing 22 species), 14% were non-susceptible to penicillin with no high-level gentamicin-resistant isolates. All β-haemolytic streptococci were susceptible to penicillin: the proportion of Group A isolates decreased from 42% in 2013 to 29% in 2017 (p<0.05).

The proportion of E. faecium among enterococci increased steadily, rising from 44% in 2013 to 50% in 2017, whereas the proportion of E. faecalis decreased from 53% in 2013 to 45% in 2017. Rates of resistance to vancomycin were 31% for E. faecium and 1% for E. faecalis.

Rates of ESBL production were higher in E. coli (10%) and Klebsiella (11%) than in Enterobacter (6%), Proteeae (0.5%) and Serratia (0.3%). Nine isolates of Enterobacteriaceae were carbapenemase-producers (2013 (n=1), 2015 (n=4), 2017 (n=4)) comprising E. cloacae (n=1, OXA-48, 2013), S. marcescens (n=2, OXA-48), Klebsiella (n=3, KPC; n=2, OXA-48, n=1, NDM).

Rates of colistin resistance in Enterobacteriaceae were 0.5% and 1% for E. coli, and Klebsiella, respectively; however, rates were higher, and increasing, in Enterobacter (6% in 2013 to 12% in 2017, p=0.03). Rates of non-susceptibility to meropenem and imipenem in Pseudomonas species were 10% and 6%, respectively, and carbapenemase-producing isolates were infrequent (n=2: VIM, NDM). Three isolates were non-susceptible to ceftolozane-tazobactam (MICs 8, 128, >256 mg/L). Only one isolate of P. aeruginosa was resistant to colistin; the isolate also had a raised MIC to carbapenems (MIC 8 mg/L meropenem) but was susceptible to ceftolozane-tazobactam (MIC 1 mg/L).

Discussion: The proportion of MRSA among S. aureus has continued to decrease since 2013, as has the proportion of Group A streptococci associated with bacteraemia. In contrast, the proportion of E. faecium has increased, which is important as this species is commonly more resistant than E. faecalis. Carbapenemase-mediated resistance among surveillance Enterobacteriaceae remains rare; however, rates of colistin resistance continue to increase in Enterobacter. Continued collection of surveillance data is crucial for our understanding of antibiotic resistance trends in the UK and Ireland.

047
BSAC Respiratory Resistance Surveillance: A review of non-susceptibility over five seasons (2012-2017)

Abstract - 047

Poster 047

BSAC Respiratory Resistance Surveillance: A review of non-susceptibility over five seasons (2012-2017)

Carolyne Horner1, Shazad Mushtaq2, BSAC Standing Committee on Resistance Surveillance1
1British Society for Antimicrobial Chemotherapy, Birmingham. 2Public Health England, London

Introduction: The British Society for Antimicrobial Chemotherapy (BSAC) Respiratory Resistance Surveillance Programme has monitored the antimicrobial susceptibility of isolates from community- (CO-) and hospital-onset (HO-) lower respiratory tract infections (LRTI) since 1999/2000 and 2008/2009, respectively. This update reviews the latest five seasons of data (October 2012 – September 2017).

Methods: Consecutive isolates causing CO-LRTI, or HO-LRTI throughout the UK and Ireland were tested, with participating laboratories (n=24-40) collecting 7-20 isolates/species. Minimum inhibitory concentrations were determined centrally by BSAC agar dilution.

Results: 11,382 isolates were tested: 5236 from CO-LRTI, comprising Streptococcus pneumoniae (n=1866), Haemophilus influenzae (n=2289), and Moraxella catarrhalis (n=1081), and 6146 from HO-LRTI, comprising Enterobacteriaceae (n=3719), Staphylococcus aureus (n=1046), Pseudomonas spp. (n=1064), and Acinetobacter spp. (n=317).

Fifty-six serotypes of S. pneumoniae were represented: 15 isolates were non-typeable. The most common serotypes were 15A (7%), 11A (6%), 3 (5%), 23A (4%) and 23B (4%). Serotype 7C, previously uncommon, was prominent (n=16) in 2016/17, largely from one area of England. Rates of non-susceptibility in S. pneumoniae were: penicillin 14%, tetracycline 16% and erythromycin 19%; 8.5% isolates were resistant to all three agents, and a third of these were serotype 15A. Three isolates, all multi-drug resistant from seasons 2012/13 and 2016/17, were resistant to penicillin, with MICs 4-8mg/L. Almost all M. catarrhalis (99%) were susceptible to all of amoxicillin-clavulanate, ciprofloxacin and tetracycline, as were 92% of H. influenzae.

Rates of MRSA among S. aureus from HO-LRTI continued to decrease (24% in 2012/13 vs. 10 % in 2016/17) and high-level mupirocin resistance, mediated by mupA, was rare (n=5 MRSA, n=1 MSSA).

Eighteen species of Acinetobacter were represented but over 50% isolates were A. baumannii. OXA-23 carbapenemase was present in 35/317 isolates whilst six had OXA-58. Thirty-two (10%) Acinetobacter isolates had ‘triple’ resistance to carbapenems, aminoglycosides and ciprofloxacin; 30 of these were susceptible to colistin whilst two, both A. baumannii, were not susceptible to any agent tested.

Six species of Pseudomonas were represented but 99% of isolates were P. aeruginosa. Rates of non-susceptibility to ciprofloxacin, piperacillin-tazobactam, ceftazidime, and gentamicin were 17%, 11%, 6%, and 5%, respectively, with no significant change over the five year period. Most isolates (99%) were susceptible to ceftolozane-tazobactam and colistin. Rates of non-susceptibility to meropenem and imipenem in Pseudomonas species were 17% and 16%, respectively, and metallo-carbapenemases were infrequent (VIM, n=4; NDM, n=1). Two isolates had extended-spectrum beta-lactamases (VEB, PER). Four isolates were non-susceptible to ceftazidime and ceftolozane-tazobactam, carbapenems, aminoglycosides and ciprofloxacin, and were only susceptible to colistin. Colistin resistance was more frequent (7%) in Acinetobacter than in Pseudomonas (0.8%).

Extended-spectrum beta-lactamases (ESBLs) were present in 11% of E. coli, 10% Klebsiella and 4% Enterobacter, whereas AmpC beta-lactamases were present in 3% of E. coli, 0.3% Klebsiella and 19% Enterobacter. Eleven Enterobacteriaceae isolates with carbapenemases were identified in four seasons (2013/14 – 2016/17) and comprised E. coli (OXA-48, n=1), E. cloacae (OXA-48, n=2) and K. pneumoniae (KPC, n=4, NDM, n=2, and OXA-48, n=2). Rates of colistin resistance excluding Proteeae, were 0.2% and 1% for E. coli, and Klebsiella, respectively, but were much higher at 7% in Enterobacter.

Discussion: Rates of non-susceptibility and common serotypes in S. pneumoniae did not change over the five year period. H. influenzae and M. catarrhalis remain largely susceptible to existing antimicrobials. The proportion of MRSA continues to decrease among S. aureus. Carbapenemase-mediated resistance is rare and most commonly seen in Klebsiella species. Colistin resistance was most common in Acinetobacter spp. and Enterobacter species. Continued collection of surveillance data is crucial for our understanding of antibiotic resistance trends in the UK and Ireland.

048
Experience of using a mobile app audit tool to support prescribing quality indicators

Abstract - 048

Poster 048

Experience of using a mobile app audit tool to support prescribing quality indicators

Jacqueline Sneddon1, Andrea Patton1, Mark Buchner2
1Heathcare Improvement Scotland, Glasgow. 2Tactuum, Glasgow

Introduction: The Scottish Government approved a national hospital antimicrobial prescribing quality indicator which was developed by the Scottish Antimicrobial Prescribing Group (SAPG) in collaboration with health board Antimicrobial Management Teams (AMTs). The quality indicator is intended to support reduction in unnecessary hospital antibiotic use through promoting review of IV therapy within 72 hours and documentation of duration for oral therapy. The data for the quality indicator are recorded via an audit tool within the SAPG Antimicrobial Companion app which was launched in August 2017.

Methods: Each month, data are collected on a minimum of 10 patients in one medical and one surgical ward on each of the following measures:

  • Indication documented
  • Compliance with local antibiotic prescribing policy
  • All prescribed doses administered
  • Oral therapy only: duration or stop date documented
  • IV therapy: Clinical review within first 72h and documented outcome of review

Data can be downloaded from the app by Antimicrobial Management Teams in each board and are collated to produce a national report.

Results: All health boards were collecting data from January 2018 onwards and data was analysed at national level. Between January and June 2018 indication for antibiotic was documented in the notes in 98% and 93% of infection episodes audited in medical and surgical wards respectively. Compliance with antibiotic policy was similar during this period for medicine (93%) and surgical (92%). For patients only receiving oral therapy, all prescribed doses were administered in 95% and 93% on average of infection episodes audited in medical and surgical wards respectively. A duration or stop date was documented in 73% and 64% respectively. All prescribed doses were administered for patients on IV +/- oral therapy in 97% and 96% of infection episodes on average in medical and surgical wards respectively. A clinical review was documented in 96% of infection episodes on average in medical wards and of the 992 clinical reviews the most common outcomes of the review were Intravenous to oral switch (IVOS) (33%), to continue on the same antibiotic with a reason documented (31%) and to continue on the same antibiotic without a reason documented (17%). In surgical wards, a clinical review was documented in 91% of infection episodes on average and of the 690 clinical reviews the most common outcomes were to continue on the same antibiotic with a reason documented (31%), to continue on the same antibiotic without a reason documented (28%) and IVOS (23%).

Discussion: Data collection via the app can be done in real-time during ward rounds and provides clinical staff with an easy to use, accessible audit tool. The data recorded on the SAPG app allows national monitoring of quality indicators as well as at local level within individual health boards. Following recent improvements to the reporting functions within the app weekly or monthly reports can be generated on any device via the app to track compliance over time and can downloaded into pdf format to email directly to clinical teams to share results. The updated version will be used from October 2018.

049
Utilisation of a national point prevalence survey of recorded penicillin allergy to inform a penicillin de-labelling process

Abstract - 049

Poster 049

Utilisation of a national point prevalence survey of recorded penicillin allergy to inform a penicillin de-labelling process

Jacqueline Sneddon1, Ronald Andrew Seaton1,2, Andrea Patton1, William Malcolm3, Neil Ritchie2
1Healthcare Improvement Scotland, Glasgow. 2NHS Greater GLasgow and Clyde, Glasgow. 3Health Protection Scotland, Glasgow

Introduction: Mislabelling of penicillin allergy is associated with sub optimal antibiotic choice, increased length of stay, increased treatment cost and poorer outcomes including Clostridium difficile infection and antimicrobial resistance. The Scottish Antimicrobial Prescribing Group (SAPG) initiated a programme of work supported by a multidisciplinary Steering Group to develop a process for de-labelling of patients with documented but unfounded penicillin allergy. As a first step an evaluation of the extent and nature of penicillin allergy recording in patients admitted to Scottish hospitals was carried out.

Methods: A point prevalence survey (PPS) of penicillin allergy labelling was undertaken using a bespoke audit tool and a draft national algorithm designed to identify patients suitable for de-labelling. Data was collected in 10 health board areas in over a one week period in each participating hospital during February and March 2018. Data were collected in admission wards focusing on newly admitted patients with a maximum of 20 patients per ward. All eligible patients were screened for a penicillin allergy label and data was collected on the nature and timing of their reaction to penicillin and any antibiotics prescribed during their current admission. Data were analysed using Microsoft Excel and results from adult wards are presented here.

Results: 20 wards were included in the PPS (8 medical admissions, 9 surgical admissions, two combined assessment units and one medicine for the elderly). 1871 patients were reviewed and 188 patients (10%) had a penicillin/betalactam allergy documented. The mean age of the patients with a documented allergy was 67 years and 64% were females. 48% of patients had an antibiotic prescribed during the current admission; a combination (excluding penicillin) was prescribed to 23% of patients, doxycycline (19%), levofloxacin (14%) and clarithromycin (10%).

Patients’ adverse reactions were associated most commonly with penicillin (52%), amoxicillin (15%) and was unknown in 12% of patients. The route was oral in 67% of patients, parenteral in 8% and unknown in 25%. The sources of information used to check details of the allergy included the patient in 58% of cases, GP record (44%), medicines reconciliation form (34%) and hospital records (24%). The reaction occurred more than 10 years ago in 62% of cases, between one and 10 years in 21% of cases, unknown in 14% and within the last year in 3%. The timing in relation to administration of the antibiotic was unknown in 59% of cases, less than one hour in 18% of cases, between one and 24 hours in 14% and over 24 hours in 10%.

Using the draft algorithm a Type 1 allergy was identified in 27 patients (14%), severe non-T1 allergy in 3 patients (2%), other adverse reaction in 40 patients (21%) and uncertain in 118 patients (70%). In Type 1 allergy patients, the main reactions to the antibiotic were reported as itchy rash (70%), swelling (33%) and breathing difficulty (26%). Other adverse reactions reported included GI symptoms (68%) and itchy rash (13%). For patients with an uncertain allergy the reaction was unknown in 47% of patients.

Discussion: The results of the PPS have allowed us to establish a baseline of prevalence, nature and impact on antibiotic treatment of a penicillin allergy label. We have now finalised an algorithm that will segment patients based on their reaction to a penicillin and identify those that can safely be offered a penicillin challenge test. A de-labelling pilot using the algorithm will be carried out in several hospitals. Data will be collected on challenges of using the algorithm, conducting the oral challenge and communicating the result including feedback from clinical teams and patients.

050
Evaluation of current practice to inform a national antifungal stewardship programme

Abstract - 050

Poster 050

Evaluation of current practice to inform a national antifungal stewardship programme

Jacqueline Sneddon1, Brian Jones2, Ronald Andrew Seaton2,1, Andrea Patton1

1Healthcare Improvement Scotland, Glasgow. 2NHS Greater Glasgow and Clyde, Glasgow

Introduction: The Scottish Antimicrobial Prescribing Group initiated a national programme to optimise prophylaxis, empirical and targeted treatment strategies for antifungal agents and to minimise unnecessary use and resistance development. The programme is supported by a multi-professional steering group including clinician representatives from Critical Care and Haemato-oncology, specialties with highest use of antifungal agents. To inform development of national consensus guidance surveys of current practice were carried out.

Methods: Surveys were used to capture information about current practice in use of antifungals in intensive care/surgery patients focused on invasive candidaemia (IC) and in Haemato-oncology patients. The ‘Smart Survey’ online tool was used and draft surveys were tested by clinicians from the Steering Group prior to dissemination. The survey link was distributed via existing clinical networks along with a pdf version of the survey to encourage a combined response from each team. Several reminders were sent about completing the surveys. Results were analysed using Microsoft Excel.

Results:

Critical care

15 responses were received from 6 health board areas. Eight units have a local guideline for IC, 6 units do not and one unit did not answer this question.

Three units never use prophylactic antifungals, 8 units use occasionally and 3 units use routinely for selected patients. Fluconazole is the antifungal agent of choice.

Three teams routinely use empirical treatment for potential/suspected IC for specific patient groups, 8 teams use it occasionally, one team use when advised by Microbiology/Infectious Disease specialist and one team was unsure. Failure to respond to antibacterials and various diagnostic tests inform the decision to start patients on empirical treatment and fluconazole is the agent of choice in most units.

For both empirical and directed antifungal treatment clinical symptoms and advice from Microbiology are the main criteria that inform discontinuation of treatment.

Eleven units would support development of national guidance to standardise practice.

Haemato-oncology

7 responses were received from 6 health board areas with at least one response from each of the 3 regional cancer networks. Five units have a local guideline for suspected fungal infection and 2 units do not.

Antifungal therapy is used to prevent fungal infections in all units and a variety of agents are used first line depending on the specific patient group.

Antifungal therapy for empirical treatment of potential/suspected fungal infections is used in neutropenic patients and decisions to initiate treatment are informed by clinical symptoms, chest X-ray and CT scan, bronchoscopy results and galactomannan testing. Caspofungin or Ambisome are used first line in most units.

For both empirical and directed antifungal treatment clinical symptoms, advice from Microbiology, recovery of neutrophils and radiology are the main criteria used to inform discontinuation of treatment.

Only four units would support development of national guidance to standardise practice in use of antifungals but comments indicated universal support for improved access to diagnostics and CT scanning to inform treatment decisions.

Discussion: There is variation in current practice in management of invasive fungal infection and use of antifungals in Critical Care and Haemato-oncology units in Scotland. Use of existing clinical networks allowed us to engage with a wide group of clinicians for each speciality area to seek their input but a limitation of both surveys was that teams in all health board areas did not participate. Survey responses have provided useful information to support development of good practice recommendations to reduce unwarranted variation. The majority of the Critical Care survey respondents were supportive of such national consensus guidance. Such guidance for Haemato-oncology requires further discussion with clinical networks.

051
Evaluating a new Carbapenemase-producing Enterobacteriaecea (CPE) screening assessment tool – can we improve screening using a single question?

Abstract - 051

Poster 051

Evaluating a new Carbapenemase-producing Enterobacteriaecea (CPE) screening assessment tool – can we improve screening using a single question?

Christopher Lander, Desmond Hsu, Maria Krutikov, Martina Cummins, Jonathan Lambourne
Barts Health NHS Trust, London

Introduction: Infections caused by Carbapenemase-producing Enterobacteriaecea (CPE) are difficult to manage clinically and are an emerging problem worldwide and also locally within UK hospitals. Identification of CPE carriage has crucial clinical and infection prevention and control implications. Both national and local risk-assessment tools used to identify patients who should be screened for CPEs use complex criteria. We investigated the utility of using a simplified screening assessment tool, consisting of a single question: “Have you had an overnight admission in the UK or overseas in the last 12 months?”

Methods: We compared ease of use and the proportion of patients who should have been screened using three risk-assessment tools; the national Public Health England (PHE) guidelines, current local guidelines and our proposed single-question screening. We also recorded the proportion of patients who were screened. Seven representative medical and surgical wards, encompassing general and specialty patients across two sites (in a London teaching hospital) were selected. A spot audit was performed on a chosen weekday. This involved examination of each patients’ electronic records and face-to-face interviews to apply the three risk-assessments.

Results: A total of 150 patients were reviewed. The proportion of patients who met the criteria for screening according to the national PHE guidelines, current local guidelines and proposed single-question tool was 49% (range 21-75%), 51% (21-75%) and 54% (21-81%) respectively. All patients who should have been screened using the PHE or current local risk-assessment would have been identified for screening using the single-question risk-assessment. Only 30% of patients who were identified as ‘high-risk’ by the current local risk-assessment tool were actually screened. Both the PHE and local risk-assessment tools were time-consuming to implement and required the patient to be able to effectively communicate in order to complete fully. 

Discussion: The current CPE screening risk-assessment tool is poorly implemented, as it is cumbersome and time-consuming. The proposed new single-question risk-assessment tool is user-friendly, simple to remember and captures all patients identified by the existing tools. The single-question tool does identify additional patients for screening and the cost implications of additional test materials and laboratory time is an important factor that should be balanced against the clinical and financial costs of a potential outbreak.

052
Safety and efficacy of Temocillin as a carbapenem-sparing agent

Abstract - 052

Poster 052

Safety and efficacy of Temocillin as a carbapenem-sparing agent

Katie Heard1, Stephen Hughes1, Nabeela Mugal1,2,3, Luke Moore1,2,3
1Chelsea & Westminster NHS Hospitals, London. 2University College London Hospital, London. 3National Institute for Health Research Health Protection Research Unit in Healthcare Associated Infections and Antimicrobial Resistance, Imperial College London

Introduction: Temocillin, a narrow spectrum penicillin-based antibiotic which provides activity against many Enterobacteriales including some AmpC and Extended Spectrum B-Lactamase (ESBL) producers. In 2016 targets to reduce carbapenem use were set in the UK via the Commissioning for Quality and Innovation (CQUIN) framework. With the ever-growing emergence of ESBL producing Enterobacteriales, temocillin may be a useful tool in reducing carbapenem use.

In 2018 the licenced dose of temocillin for severe infections was increased to 6g/day in normal renal function, possibly demonstrating better outcomes however with significant financial costs if implemented across the board.

To understand current experiences at 4g/day dosing, we undertook a single centre retrospective observational study of the outcomes of patients prescribed temocillin.

Methods: All adult inpatient electronic prescriptions of temocillin (3 days or greater) from March 2016 to July 2017 were retrieved using a clinical decision support system (ICNET®). ICNET® was then used to retrieve patient demographic and clinical data, including mortality, readmissions, need for antimicrobial escalation, episodes of Clostridium difficile, occurring within 30 days of temocillin ending.

Treatment success was defined as survival, no use of broad-spectrum agent for the same indication, no subsequent readmission for infection, and no C. difficile, occurring within 30 days of completing temocillin.

Treatment was considered empirical if the patient had no positive microbiology results or if results were greater than 30 days old.

Results: Temocillin was used in 148 patient-episodes during the study period. The average patient age was 76.6years (range 20-102years), 45% were female. Pre-temocillin average length of stay was 16.3days (range 0-156days). Prior to temocillin treatment 7 patients were identified as C. difficile carriers. Average temocillin course length was 7.3days (range 3-42days). Indications for use: urinary tract infection (UTI) 62.8% (n93), pneumonia 26.3% (n39), pneumonia/UTI 2.7% (n4), unlicensed sites: skin 2.0 (n3), prostate 1.4% (n2) and abdomen 4.7% (n7). 110 patients had targeted treatment; 57 patients with Escherichia coli infection, 53 with other Enterobacteriales. 38 patients were treated with temocillin after failing first line antimicrobials. All patients were dosed at 2g BD; in cases of reduced renal function doses followed local guidance.

Overall temocillin treatment success was 75.6%, highest when used to treat UTIs (82.8%, OR 2.75, CI:1.27-5.94, p=0.01). A non-significant difference in success was observed between patients with E. coli and other Enterobacteriales infections (82.5% vs 69.8% p=0.18). Empirical treatment demonstrated a 73.7% success rate (vs 76.4% among targeted treatment, p=0.92). In terms of the treatment failures;

  • There were no incidences of C. difficile within 30 days of temocillin treatment, 1 patient had C. difficile negative diarrhoea. There were no incidents of fits or other severe adverse reactions reported.
  • 30day mortality rate was 8.8% (n13). There was no significant difference between those that were being treated for pneumonia vs UTI (15.8% vs 6.5%, p= 0.18).
  • 27 patients were treated with a broad-spectrum antimicrobial within 30days of receiving temocillin, 11 for a different indication. Of the 16 escalated for the same indication, 4 received an antipseudomonal (ceftazidime or piperacillin-tazobactam), 2 ceftriaxone, 9 meropenem and 1 had vancomycin added. The most common reasons for escalation were non-responding pneumonia (31.2%) or in response to new microbiology (18.8%). Of the 5 patients that had therapy changed due to non-responding pneumonia 4 died.

Discussion: In this single study retrospective review, Temocillin (4g/day) provides an effective and safe antimicrobial treatment option for treatment of Gram negative infections. This overall cure rate should be considered in the context of the age and significant co-morbidities of these patients. Increased dosing may be indicated where bacteria are known to have high break-points (further research required) or when the infection is not of a urinary source.

053
Secrets of the hospital underbelly: metagenomics of AMR in hospital sewage

Abstract - 053

Poster 053

Secrets of the hospital underbelly: metagenomics of AMR in hospital sewage

Meghan Perry1,2, Bram Van Bunnik2, Luke McNally2, Frank Aarestrup3, Mark Woolhouse2
1NHS Lothian, Edinburgh. 2University of Edinburgh, 3Technical University of Denmark, Copenhagen

Introduction: An important aspect of research into antimicrobial resistance is understanding its ecosystem. It is well established that hospital wastewater can contain multidrug resistant organisms but no previous attempt has been made to correlate the resistance in sewage with antimicrobial usage within the hospital. This work aims to study whether resistance gene abundance levels in sewage can reflect clinical activity within the hospital.

Methods: Using composite samplers sewage was collected over a 24 hour period from 8 collection points at the Western General Hospital, Edinburgh and simultaneously from community sewage works. High throughput shotgun sequencing was performed using Illumina HiSeq4000 and reads analysed using the online pipeline MGmapper and Resfinder. The results were analysed with databases of hospital antimicrobial usage and clinical isolates using R.

Results: Each collection point, as a representation of a range of medical and surgical specialties, had different patterns of antimicrobial use. There was variance in resistance gene abundance and diversity in sewage between each collection point and between hospital and community sewage samples. Using a generalised liner mixed effects model accounting for random effects of observation, collection point and antimicrobial class, we saw a clear relationship between both length of stay and antimicrobial usage at the class level on resistance gene abundance in hospital sewage. No relationship between resistance in clinical isolates and resistance gene abundance in sewage was found.

Discussion: This is the first representation of how resistance gene abundance in hospital sewage reflects hospital antimicrobial usage and demonstrates that antimicrobial usage is a major driver of antimicrobial resistance gene outflow from the hospital into the sewage environment. It further emphasises the importance of minimising antimicrobial use to minimise faecal carriage of AMR and may have implications for infection control.

054
An audit of admission screening for carbapenemase-producing Enterobacteriaceae

Abstract - 054

Poster 054

An audit of admission screening for carbapenemase-producing Enterobacteriaceae

Matthew Powell1, Edward Bevan1,2, Katie Hardy2
1University of Birmingham. 2University Hospitals Birmingham NHS Foundation Trust

Introduction: Carbapenemase-producing Enterobacteriaceae (CPE) represent a challenge to infection control specialists and are a significant risk to public health. In 2013 Public Health England introduced a CPE toolkit, which provides strategies to prevent the spread of CPE within healthcare settings. The toolkit recommends that patients with suspected CPE colonisation or infection are isolated and screened on admission, and that three rectal swabs are taken 48 hours apart if the initial screens are negative. This study aimed to assess compliance with recommendations from this toolkit at University Hospitals Birmingham (Heartlands Hospital).

Methods: We undertook a retrospective audit of patients previously screened for CPE in 2017 to determine (1) The proportion of CPE positive patients who were screened and isolated on admission; (2) The number of consecutive rectal swab cultures required to identify CPE in carriers and (3) Adherence with the above rectal screening protocol for patients in whom initial screening is negative.

Results: Three hundred and thirty one patients were screened for CPE carriage of which 72 were CPE-positive and 259 CPE-negative. 81% of CPE-positive patients were screened and isolated on admission. 83% of CPE-positive patients were identified by the first rectal screening swab. 39% of CPE-negative patients were correctly screened with rectal swabs in line with the toolkit protocol as described above.

Discussion: Compliance with the CPE toolkit recommendations for isolation is generally good for CPE-positive patients but is lower for obtaining three negative screens in CPE-negative patients. We have shown that the number of rectal swabs used for screening could be reduced, which would result in a significant cost saving for the trust and would allow de-escalation of isolation precautions after only a single negative CPE rectal screen.

055
Antimicrobial prescribing – lessons from a Stroke Ward

Abstract - 055

Poster 055

Antimicrobial prescribing – lessons from a Stroke Ward

Michael Sawaryn, Jonathan Moore
Gateshead NHS Foundation Trust

Introduction: The incidence of sepsis is increasing in the developed world, and hence the use of IV antibiotics will continue to be a common therapy administered in hospitals, necessitating conscientious, evidence-based IV antibiotic use —’antibiotic stewardship’. The UK government has committed to improve antimicrobial stewardship by developing a Commission for Quality and Innovation (CQUIN) target aiming to reduce antimicrobial resistance by reducing dependence on broad-spectrum agents and improving timely antibiotic review. The Gateshead Health NHS Foundation Trust updated their Antimicrobial Prescribing guidelines in 2017 following the updated CQUIN targets. We conducted this audit to quantify improvements in antibiotic prescribing practices and then went on to address the challenges of improving antibiotic reviews by asking the major prescribers in our hospital—Foundation doctors—for their opinions on making changes in this area.

Methods: Data was collected from patient notes on a Stroke/Care of the Elderly ward at the Queen Elizabeth Hospital, Gateshead of those who were, or had been, on IV antibiotic therapy during that admission for clinically suspected infection. Group 1 data was obtained between April and July 2017, and a Group 2 data between January and April 2018. ‘Appropriateness of IV antibiotic prescribing’ and ‘Review within 72 hours’ were collected for Group 1. ‘Initial review date’ and ‘Seniority of antibiotic review’ (i.e. ST3 registrar or above) were also included for Group 2. Finally, a qualitative survey establishing opinion on current antibiotic review practices was carried out, asking Foundation Doctors in an open-floor session 1) why antibiotics are not reviewed; and 2) what could be done to improve antibiotic review.

Results: Overall, 86.7% (Group 1, n=15) and 81.3% (Group 2, n=16) of IV antibiotics were prescribed in accordance with Trust Antimicrobial guidelines, or had clearly documented justification for going against the guidelines. 80% (Group 1) and 93.8% (Group 2) of antibiotic prescriptions were reviewed within 72 hours. 12.5% of Group 2 patients had an initial review date documented at the point of prescribing. 81.3% of antibiotics reviews were performed by senior medical staff. The qualitative survey demonstrated some of the barriers to reviewing antibiotics and found that the most popular proposal for improving antibiotic reviews would be to seek electronic prescribing solutions. 

Discussion: IV antibiotics prescriptions were generally appropriate, however this area requires improvements—the authors suggest a target of 95% agreement with available guidelines or adequate documented justification for prescribing against guidelines. The initial review date was poorly documented generally, and is an area for improvement, and may improve ongoing patient care as well as inter-team communication. The majority of prescriptions were reviewed within 72 hours, however this number was lower when accounting for seniority of review. Furthermore, as these rates were similar to those achieved by the Trust as a whole, hospital-wide, system changes should be made in order to achieve the CQUIN target of 90% of antibiotics reviewed in 24-72 hours. A popular option to improve antibiotic reviews is to implement or develop electronic prescribing systems to give reminders to clinicians when antibiotics are due to be reviewed. 

056
Drivers for meropenem prescription: a retrospective study at a large district hospital, UK

Abstract - 056

Poster 056

Drivers for meropenem prescription: a retrospective study at a large district hospital, UK

Aimee Goel, Zhilin Jiang, Tejal Vaghela, Hala Kandil
West Hertfordshire Hospitals NHS Trust, Watford

Introduction: Meropenem is often a last resort antimicrobial, and its use must be carefully monitored and limited. At West Hertfortshire Hospitals NHS Trust (WHHT), we noted the increase in meropenem consumption from 47.9 to 76.1 DDD/1000 admissions over 3-month period from January to March 2018. The aim of this audit was to identify the drivers of meropenem prescription at our hospital, and to develop a strategy to reduce inappropriate meropenem use.

Methods: A list of all March 2018 adult inpatient prescriptions for Meropenem was obtained from the pharmacy system. Admissions summaries, clinical notes and microbiology reports were analysed to identify the indication for meropenem prescriptions. Prescriptions were regarded as appropriate if one or more of the following criteria was met:

  1. Non severe penicillin allergy
  2. Pathogen/susceptibility guided therapy
  3. Clinical indication for escalation to broader spectrum antibiotic

Results: Thirty seven patients were included in our study. Aspiration pneumonia, urosepsis and urinary tract infection were the main indications for the use of meropenem in these patients. Fifty five percent of meropenem prescriptions were found to be as a result of escalation of antibiotic therapy, of which 81% were escalated from piperacillin/tazobactam. The clinical indication or microbiologist approval for this escalation was not documented and hence was difficult to assess appropriateness of meropenem use in some of these cases. 

Fourty five percent of prescriptions were guided by pathogen and antibiotic susceptibility with 42% of the pathogens recovered being ESBL producers, while 24% of patients who received meropenem had documented non-severe penicillin allergy.

The increase of meropenem consumption was also attributed to prescription for duration longer than the standard recommended duration for the clinical indication.

Discussion: Overall, 66% of meropenem prescriptions at WHHT were appropriate. The main drivers were multi-resistant pathogens and documented non severe penicillin allergy.

Further study is needed to investigate the indication for the escalation to meropenem and the indication for prolonged meropenem courses.

Introduction of a robust real-time monitoring of meropenem is needed to ensure appropriateness of meropenem prescription.

057
Appropriateness of antibiotic use in hospitalized patients: a retrospective study at a large district hospital, East of England, UK

Abstract - 057

Poster 057

Appropriateness of antibiotic use in hospitalized patients: a retrospective study at a large district hospital, East of England, UK

Farhana Ahad, Trishan Patel, Mehreen Paraouty, Tejal Vaghela, Hala Kandil
West Hertfordshire Hospitals NHS Trust, Watford

Introduction: Effective antibiotics are becoming increasingly limited and inappropriate antibiotic prescribing is a growing global health concern. Hence, it is essential that the use of antibiotics is monitored in order to preserve their future effectiveness. This audit was undertaken to assess the appropriateness of antibiotic therapy in adult patients within the acute hospital setting.

Methods: Data was collected over a period of 2 months. Assessment of probable infection was made by reviewing patient observations and sepsis criteria, consideration of the biochemical inflammatory markers (white cell count and C-reactive protein) and analysis of microbiological and radiological findings. Information of patients’ medical history (including working and final diagnosis) and the choice and length of antibiotic was included within the data.

Results: There was a total of 77 courses of antibiotics across 45 patients. The indications for these antibiotics included 31% of cases with respiratory tract infections (RTI), 27% with urinary tract infection (UTI), 18% of intra-abdominal infection, 9% of skin infections, 7% with infection of unknown source and 9% of other causes. A total of 86% of the antibiotics were recognized to have a justified indication for use, and 98% of these courses were deemed necessary and not redundant. However, despite being indicated, 10% of these antibiotics were the incorrect antibiotic according to local guidelines. Of the courses of antibiotics not indicated, 14% of these cases were antibiotics prescribed for infections with an unconfirmed source, 43% for unconfirmed UTI and 43% for clinically unconfirmed RTI. Moreover, 47% of patients were noted to have one or more non-essential days of antibiotic therapy, with 43% of these being due to an antibiotic not being indicated. 10% of patients had unexplained continuation despite infection being ruled out, 29% with unexplained continuation beyond standard duration and 19% with explained continuation beyond standard duration.

Discussion: Our findings from this preliminary collection of data show that in the majority of cases, antibiotics are appear to be appropriately prescribed. However, there are still number of cases whereby antibiotics are wrongly or inappropriately prescribed or continued excessively, which may impact the efficiency of these antibiotics for future infections in these patients.

058
Pre-hospital antibiotics administered by paramedics improve outcomes for patients with septic shock

Abstract - 058

Poster 058

Pre-hospital antibiotics administered by paramedics improve outcomes for patients with septic shock

Jon Chippendale1, Bethan Stoddart2
1Clinical Development Lead, East Midlands Ambulance Service. 2Lead Consultant Microbiologist, Path Links, North Lincolnshire and Goole NHS Trust

Introduction: Septic shock is a medical emergency necessitating urgent antibiotic and supportive treatment. It has been shown that delaying effective antibiotic administration has a detrimental impact on survival. Paramedics are ideally placed to deliver antibiotics to patients with septic shock, as they are frequently the first healthcare professionals with whom they have contact.

A pilot project undertaken by East Midlands Ambulance Service (EMAS) within Northern Lincolnshire showed that paramedics can accurately identify patients with septic shock, initiate the sepsis six including undertaking blood cultures, and can safely administer a broad spectrum antibiotic within a closely defined protocol. During the initial pilot, there was a significant reduction in time from identification of sepsis to administration of an antibiotic. Following roll-out to the whole of Lincolnshire in December 2017, this study analyses the patients identified as having septic shock by EMAS, comparing the 30 day mortality and other outcomes between the group receiving pre-hospital antibiotics and those whose antibiotic treatment was delayed until their arrival and assessment with hospital A&E departments.

Methods: The East Midlands Ambulance Service Medusa database was interrogated to identify all patients between 1st December 2017 and 26th June 2018 assessed by a paramedic to have sepsis and who had a systolic blood pressure <90mmHg. Ambulance and laboratory records, electronic discharge summaries and patient administration systems were viewed to establish demographics (age, gender, admitting location, length of stay), laboratory results including positive and contaminated blood cultures, survival to and diagnosis at discharge, and mortality at days 0, 7, 30 and 90.

Results: Patients presenting to EMAS with septic shock and admitted to a Lincolnshire Hospital (NLG or ULH) = 252; of those 65 received pre-hospital antibiotics.

Blood culture significant organism positivity rates

            Prehospital antibiotic group     24.6%

            No pre-hospital antibiotics       15.2%

Thirty day mortality available for the first 186 records analysed (we are awaiting the records from the remainder and these will be included these in the final analysis).

            Prehospital antibiotic group     26.9%

            No pre-hospital antibiotics       30.6%

            Overall 30 day mortality          29.5%

Selected subgroups 30 day mortality

            Likely chest focus

                        Prehospital antibiotic group     11.8%

                        No pre-hospital antibiotics       32.6%

            Likely Gram negative sepsis (UTI / abdo focus)

                        Prehospital antibiotic group     17.6%

                        No pre-hospital antibiotics       28.1%

The impact on length of stay, blood culture results, subgroups, 7, 30, 90 day mortality and survival to discharge data will be presented in detail.

Discussion: Prehospital antibiotics improve outcomes for patients with septic shock surviving the first 24 hours. In particular there is a reduction in 30 day mortality in patients presenting with sepsis of chest origin, those with Gram negative sepsis including abdominal and urinary sepsis, and those with true positive blood cultures. The pilot study design was not powered to include clinical outcomes but was intended for proof of process. Because of small numbers, statistical significance at p<0.05 has not been reached for overall thirty day mortality, however it has been reached for other parameters. There have been concerns expressed about the antibiotic stewardship elements of this initiative, and we have responded by restricting the original protocol which included all with ‘red flag’ sepsis to only those with septic shock. We hope that this initiative can be rolled out to benefit all patients presenting to EMAS with septic shock across the East Midlands, as we believe it to be genuinely life-saving.

059
Choosing Wisely – a synthesis of international English language infection recommendations

Abstract - 059

Poster 059

Choosing Wisely – a synthesis of international English language infection recommendations

Alexander Richards1, Gavin Barlow1,2
1Hull and East Yorkshire NHS Trust, Hull. 2Hull York Medical School

Introduction: Choosing Wisely is a global health education campaign started by the American Board of Internal Medicine in 2012, aimed at reducing unnecessary testing or treatment. The campaign brings together information from a wide range of medical institutions in the United States (US) from a variety of specialities to improve doctor-patient communication and tackle the overuse of medical resources. Based on work done in the late 1990’s and early 2000’s, data suggests that up to 30% of health-care spending within the US is wasteful, and that the inefficiency appeared to be physician driven. 

Since the initial work in the US, a number of countries have taken up this initiative, but the information and recommendations generated by these independent campaigns has not been collated. This project aims to synthesise existing international Choosing Wisely recommendations relevant to infection. 

Methods: A search of the internet was undertaken for published English language (under the banner of Choosing Wisely) recommendations relating to the investigation and treatment of infection. Identified recommendations were extracted and organised within an Excel spreadsheet by: country of origin; the organisation developing the recommendation(s); the recommendation(s) and presented supporting rationale; and the related speciality. Data was then organised thematically to highlight areas of agreement/similarity between Choosing Wisely recommendations.

Results: Recommendations from USA, Canada, United Kingdom (supported by BSAC), Wales (though no recommendations related to infection were listed), Italy, Australia and New Zealand were identified. Campaigns without an online English language translation were also identified from Brazil, Germany, France, Switzerland, Japan and South Korea, but have not been analysed at this stage. A total of 163 infection related recommendations were identified across all 7 countries. The largest proportion (19%) of recommendations related to viral upper respiratory tract infections and the use of unnecessary antibiotics. The next largest group (17%) were related to the urinary tract and reducing the unnecessary investigation and treatment of asymptomatic bacteriuria. Also of note are the recommendations (11%) relating to the testing and treatment of Clostridium difficile infection in patients without evidence of diarrhoea. 

There appeared to be a relatively high level of broad theme agreement between different countries; for example, urinary tract infections show 6 out of the 7 countries aligned with 95% of the recommendations being similar; for upper respiratory infections 6 of the countries aligned with 78% of the recommendations showing good agreement; and for C. difficile infection 4 of the countries there is good correlation (55%) of the recommendations.

In addition 18% of recommendations focused on reducing testing, including daily monitoring of inflammatory markers and serial chest x-rays to check the response to antibiotic treatment. Also 55% of the recommendations referenced the appropriate de-escalation of antibiotics making it an important theme of the studies.

Discussion: Although the campaigns were from different countries with a variety of healthcare systems, there was relatively good alignment of recommendations. The opportunity to optimise the testing and therapy, and the cost-effectiveness of care, of infections, therefore appears to be similar across high income countries. Ultimately, the success of such campaigns will depend on healthcare professionals and patients supporting recommendations and evidence to demonstrate a beneficial impact of doing so. There is currently a lack of Choosing Wisely recommendations in low and middle income countries.

060
“On-the-go” real time meropenem database: innovative approach to microbiologist advising with a positive impact on meropenem stewardship at a large district hospital

Abstract - 060

Poster 060

“On-the-go” real time meropenem database: innovative approach to microbiologist advising with a positive impact on meropenem stewardship at a large district hospital

Robert Gray, Tejal Vaghela, Rakan El-Hamad, Madhuri Vidwans, Hala Kandil
West Hertfordshire Hospitals NHS Trust, Watford

Introduction: It was noted that meropenem use at a large district hospital was steadily rising over the past 6 month period despite monitoring its use. We previously demonstrated (1) that microbiologists may themselves need to monitor and audit their antibiotic advising practice. In order to ensure that meropenem use, including those which were prescribed based on microbiology advice are appropriate and closely monitored we introduced an easily accessible, on-the-go, secure database to allow real time sharing information and monitoring meropenem use.

Methods: A secure encrypted database was introduced at West Hertfordshire Hospitals NHS trust on July 6th 2018. This database is accessed by all microbiologists and the antimicrobial pharmacist. When meropenem is advised by a microbiologist, this information goes on the system with all the details about the indication, the duration and the stop/review date. This information gets notified in real time to all other microbiologists and the antimicrobial pharmacist. The information was then used to stop/review meropenem in a timely fashion, with the data disseminated to the respective ward pharmacist. A pilot study was conducted over a period of 5 weeks, and the data was analysed.

Results: Over the 35-day period there were 61 patients prescribed meropenem. Of these 38 were approved by microbiologists (62 percent).

Over the study period the total measured consumption of meropenem by the trust decreased: the defined daily dose for meropenem in July was 52.5 per 1000 patient days, which was down on the averaged figure of 82.1 for the three preceding months, a 26% reduction.

Discussion: The reduction in meropenem use over this pilot study shows this approach is a promising innovation for improving the antimicrobial stewardship in this trust with relation to meropenem prescribing. This innovation has also contributed to changing microbiologist behaviour in relation to advising effective alternative therapy and preserving meropenem.

Further parallel studies are now required to see whether this reduction is sustained over a longer period as well as to understand the indications driving meropenem prescribing by clinicians to allow these to be targeted.

061
Collation of antimicrobial guidelines across Yorkshire and the Humber: opportunities for practice improvement

Abstract - 061

Poster 061

Collation of antimicrobial guidelines across Yorkshire and the Humber: opportunities for practice improvement

Avril Lynch1, Kevin Frost2, Stuart Bond3, Jade Lee-Milner3, Kathryn Aston3
1Sheffield Teaching Hospitals NHS Foundation Trust. 2Airedale NHS Foundation Trust, Keighley. 3Mid Yorkshire Hospitals NHS Trust, Wakefield

Introduction: The Yorkshire and Humber antimicrobial pharmacy (YAHAP) group consists of antimicrobial pharmacists and technicians from 13 acute trusts within the region. The group works together to help improve antimicrobial stewardship and infection management through networking, sharing best practice, peer-support, education and joint initiatives. YAHAP group is accountable to the Yorkshire and Humber Chief pharmacists group.

 

The British Society of Antimicrobial Chemotherapy (BSAC) recommendations for antimicrobial stewardship involve a three pillar approach to overcoming antimicrobial resistance (AMR), the first being optimising the use of existing antimicrobial agents1. Production of antimicrobial guidelines can help to optimise use of antimicrobial agents. NICE2 and BSAC1 state that empirical antimicrobial guidelines should be based on relevant national recommendations but also local microbiology, epidemiology and antimicrobial susceptibility. Another driver of changes to antimicrobial use is the NHS England AMR CQUIN3. Due to the similar geographical locations of the YAHAP trusts, the antimicrobial resistance trends for common infections are comparable and hence antimicrobial guidelines can be compared.

 

Therefore the YAHAP group decided to collate their local antimicrobial guidelines for urinary tract infections, community acquired pneumonia (CAP) and hospital acquired pneumonia (HAP), to review for trends and anomalies between the trusts.

Methods: Over a 6 month period, representatives from all the regions trust submitted their current local guidelines to a nominated member of the group to collate the antimicrobial guidelines for these specific infections of CAP, HAP and urinary infections. These are all common infections, which all the trusts had guidelines for and are regularly prescribed. As expert pharmacists within this field, we begun to analyse for trends and anomalies between the regional and national guidelines e.g. NICE and the British Thoracic Society (BTS) for CAP treatment.

Results: Antibiotic agents, routes and durations varied between the YAHAP group guidelines for CAP, HAP and urinary tract infections. The recommendations for broad spectrum intravenous antibiotics such as piperacillin/tazobactam differed between the trusts, with some more actively moving away from these agents, a drive influenced by good antimicrobial stewardship practice, the AMR CQUIN and supply issues. Variation in de-escalation guidelines and intravenous to oral switches were also apparent between the trusts guidelines.

Discussion: The collation of antimicrobial guidelines from the Yorkshire and Humber area has allowed individuals in our region to benchmark their trusts against others and highlight potential antimicrobial guidance for review. It has facilitated a critique of national and local guidelines considering local resistance patterns. This allows production of exemplar antimicrobial guidelines for review among antimicrobial stewardship teams, to facilitate optimum patient care in our region.

 

Standardisation between our trusts, to form a single antimicrobial guideline throughout the region, could aid junior doctor’s adherence to these guidelines, as they often move between the trusts and get confused between the differing recommendations. Challenges may arise in engaging local microbiologists and specialty teams to adopt a regional guideline and antimicrobial stock issues could be exacerbated if all the trusts adopted a single treatment pathway for these infections.

References

  1. British Society of Antimicrobial Chemotherapy. BSAC: Practical Guide to antimicrobial stewardship in hospitals. http://bsac.org.uk/wp-content/uploads/2013/07/Stewardship-Booklet-Practical-Guide-to-Antimicrobial-Stewardship-in-Hospitals.pdf (accessed 9th August 2018).
  2. National Institute for Health and Care Excellence. NICE: Antimicrobial stewardship: Prescribing antibiotics. KTT9. https://www.nice.org.uk/advice/ktt9/resources/antimicrobial-stewardship-prescribing-antibiotics-pdf-1632178559941 (accessed 9th August 2018).
  3. NHS England Antimicrobial Resistance 2017/19 CQUIN. https://www.england.nhs.uk/nhs-standard-contract/cquin/cquin-17-19/ (accessed 10th August 2018)

062
Glycopeptide optimisation: a patient safety approach at a large teaching hospital

Abstract - 062

Poster 062

Glycopeptide optimisation: a patient safety approach at a large teaching hospital

Kelly Atack, Philip Howard, Abimbola Olusoga, Caroline Walker, Joseph Spencer
Leeds Teaching Hospitals NHS Trust

Introduction: In 2015, a review of antimicrobial Datix incidents identified 51 incidents involving intravenous vancomycin, with some causing serious harm. Errors included incorrect or missed doses and plasma concentration levels not being taken correctly or reviewed appropriately despite a specific chart and written guidance available in the trust. Two Cochrane reviews in 2009 showed that teicoplanin was as effective as vancomycin, but caused less harm. To reduce glycopeptide errors, most vancomycin was switched to linezolid first-line where appropriate, or teicoplanin in most cases in adults initially.

76% (22) of adult antimicrobial treatment guidelines containing intravenous vancomycin were switched, with the exception of central nervous system and prosthesis infection guidelines. Linezolid guidelines were written with monitoring requirements and drug interactions to avoid to reduce the risk of serotonin syndrome. Teicoplanin guidelines were written for initial dosing depending on renal function and indication, when to take plasma concentration levels, target levels and appropriate dose adjustment. An electronic prescription protocol for teicoplanin was developed, which calculated doses appropriate for renal function and highlighted when levels were due to.

After the switch on 18th September 2017, an audit was undertaken to identify whether linezolid and teicoplanin were prescribed and monitored as per the new guidance.

Methods: The pharmacy dispensing system JAC was used to identify patients prescribed linezolid and teicoplanin between September and November 2017. The trust had both paper and electronic medication throughout this time, so both were reviewed charts to identify whether linezolid/teicoplanin were prescribed as per guidelines.

The electronic patient record (PPM+) was used to review results and identify whether patients had teicoplanin levels taken at the correct time and whether they were in the range.

A re-audit was then undertaken in August 2018.

Results: 100% (48) of patients prescribed linezolid were prescribed as per guidelines. 35% (17) of patients were prescribed intravenous linezolid, of which most were indicated for sepsis. 29% (14) of patients were prescribed linezolid for >7 days, with 100% (14) of these having the required weekly FBCs.

Between September and November, 60 patients prescribed teicoplanin were reviewed, with 42% (25) having the correct loading and maintenance dose prescribed. This increased to 90% (18) in August 2018

In August 2018, 100% (9) of patients who required a level had this taken at the correct time compared to 30% (18) at baseline, with 63% (5) of these being in range versus 32% (19) at baseline. Appropriate dose amendment occurred for 100% (4) of patients in August, versus 50% (9) between September and November.

Discussion: Since the switch, to date 14 incidents have been reported regarding incidents with vancomycin use in adults with none causing serious harm. Linezolid prescribing appears appropriate, though clinicians need more education to prescribe this orally and to reserve intravenous preparations for sepsis or patients with absorption issues.

Teicoplanin prescribing and monitoring has significantly improved since the switch. Electronic prescribing has now been fully rolled out across all adult wards and the use of the protocol has been influential in this improvement. Further education needs to be undertaken to ensure that the electronic prescribing protocol is used to achieve correct dosing and dose banding of teicoplanin as per guidelines, ensuring levels are taken on the correct day and are subsequently reviewed.

063
Interrupted time series analysis of an antimicrobial stewardship intervention to optimise prescribing for healthcare associated pneumonia

Abstract - 063

Poster 063

Interrupted time series analysis of an antimicrobial stewardship intervention to optimise prescribing for healthcare associated pneumonia

Jennifer Elias1, Kieran Lall1, Holly Lamont1, Stephen Hughes1, Luke Moore1,2,3, Katie Heard1
1Chelsea & Westminster NHS Hospitals, London. 2National Institute for Health Research Health Protection Research Unit in Healthcare Associated Infections and Antimicrobial Resistance, Imperial College London. 3North West London Pathology, London

Introduction: Healthcare associated pneumonia (HAP) is a significant cause of morbidity and mortality in UK hospitals. Excess antimicrobial prescribing for HAP is common; in particular unindicated use of antipseudomonal agents can be associated with increased antimicrobial resistance and Clostridium difficile infection. External drivers for change, such as the Commissioning for Quality and Innovation (CQUIN), incentivise Trusts to promote antipseudomonal sparing regimens. In April 2017 the study Trust changed the first line antimicrobial for HAP from an antipseudomonal agent (piperacillin-tazobactam/ciprofloxacin for beta-lactam free) to a non-antipseudomonal agent (co-amoxiclav/doxycycline for beta-lactam free), unless the patient had grown a previous Pseudomonas spp. or was considered high risk for this organism (namely; recent level 2 care, structural lung disease, severely immunocompromised). To investigate the impact of these changes to empirical treatment for HAP on patient outcomes and prescribing, a single centre retrospective observational study was undertaken.

Methods: Adult inpatient electronic prescriptions for antimicrobials for the indication of HAP were identified during two months pre- (July-August 2016) and two months post-intervention (July-August 2017). Patients with freetext indications stating ‘HAP’ or a respiratory related infection were screened and included if the Trust diagnostic criteria for HAP were met. Clinical parameters were collected via the electronic patient management system including CRP, chest radiograph, antimicrobial choice (including any antipseudomonal escalation and intravenous to oral stepdown), length of hospital stay, 30-day mortality, and readmission within 30-days. 10% of each data set was cross-validated between investigators.

Results: A total of 36 patients with HAP were identified from the study months in 2016 (mean age= 75 years (30-101years); 43% female) and 81 in 2017 (mean age=78years (36-99years); 49% female). Patient length of stay (LoS) pre-treatment for HAP was similar between 2016 and 2017 (21days vs 27days; p=0.26). Both cohorts had an identical mean Charlson Co-morbidity index (6) and >90% of patients in each cohort underwent diagnostic imaging before commencing antibiotic therapy. Of those with imaging 61% of patients had radiological evidence of HAP (defined as new infiltrates) on chest radiograph in 2016 and 62% in 2017 (p=1.0). Mean CRP prior to treatment was 123mg/L (3-386mg/L) in 2016 and 119mg/L (3-383mg/L) in 2017) (p=0.98).

Regarding impact on prescribing; the 2016 cohort received a mean of 6days (1-18days) of intravenous therapy vs 5days (0-22days) in 2017, and total course length mean was 8days in both study cohorts (1-23 vs 1-22; p=0.84). Empirical antipseudomonal usage was reduced following the intervention (81% of prescriptions in 2016 vs 36% in 2017; p=<0.01). In 2017 20 (24%) patients were considered as having/being at risk of Pseudomonas spp., of which 5 received an antipseudomonal first line and 6 of the remaining 15 subsequently required escalation to an antipseudomonal. This was compared to 8/59 (14%) of those who were assessed as not having/low risk of Pseudomonas spp. who needed escalation to an antipseudomonal agent in 2017.

Regarding impact on patient outcomes, there was non-inferiority with the new regimes, and a trend towards improvement. Specifically, post HAP: length of inpatient stay fell (2016 51 days vs 2017 37 days;p=0.22), 30-day mortality fell (2016 26% vs 2017 20%;p=0.51), and 30-day readmission fell (2016 30% vs 2017 6%;p=0.07).

Discussion: Overall, the change to empiric HAP prescribing guidelines decreased use of antipseudomonal agents and increased antimicrobial prescribing heterogeneity, with no significant negative impact on patient outcomes. Note is made in the difference in treatment numbers between 2016 and 2017, despite the same methods have been used; admission data has been assessed in each group and currently no justification for the difference in diagnosed HAP can be identified.

064
Review of the application and outcomes of the protocol for the management of pre-hospital sepsis with paramedic diagnosis and delivery of antibiotics at Isle of Wight NHS Trust

Abstract - 064

Poster 064

Review of the application and outcomes of the protocol for the management of pre-hospital sepsis with paramedic diagnosis and delivery of antibiotics at Isle of Wight NHS Trust

Emily Macnaughton, Mel Stevens
Isle of Wight NHS Trust, Newport

Introduction: A protocol for administration of intravenous piperacillin/tazobactam by paramedics before arrival at hospital was introduced at the Isle of Wight NHS Trust in 2013 to improve delivery of the sepsis care bundle, particularly time to antibiotic administration. Following successful implementation of the protocol for patients with neutropenic or post chemo sepsis and catheter associated urinary tract infection related sepsis, the protocol was broadened in 2015 to include most adults with suspected sepsis. Previous work has shown that paramedics could accurately recognise sepsis, aseptically take blood cultures and safely and rapidly commence treatment before reaching the emergency department (ED).

We undertook a review of patient characteristics and outcomes for a cohort of 30 patients who were administered pre-hospital antibiotics under the protocol in January 2018 in order to understand the patient population in whom it is used, assess appropriateness of the current antibiotic used and to inform future protocol updates.

Methods: The first thirty patients who received the protocol from 1/1/2018 were included in this review of the paramedic database, laboratory results and electronic discharge summaries. Paramedic sepsis diagnosis was compared with subsequent ED or admission discharge summary diagnoses, as recorded in the patient electronic record. Duration of any subsequent hospital admission was ascertained and 30 day mortality reviewed. Blood culture positivity and contamination was identified, along with assessment of antibiotic susceptibility to piperacillin/tazobactam of significant pathogens isolated from samples taken within 72 hours of admission.

Results: Demographics of the population given pre-hospital piperacillin/tazobactam were as follows,

Male 13 (43%), Female 17 (57%)

Age range (years): Males 35 – 89 (mean 67) Females 44 – 100 (mean 80)

Preliminary results indicated agreement between paramedic and discharge diagnosis of sepsis in 29 of 30 patients (96.7%). Mean length of stay was 6 days (range 2 -30), and 10 patients were discharged without admission from the emergency department. Two patients of 30 episodes died; 30 day mortality was 6.7%.

All 30 patients had blood cultures taken and 8 patients had positive cultures, of which 6 grew likely skin contaminants, giving a contamination rate of 20%. One of the contaminated samples also grew a Group B beta haemolytic Streptococcus, a likely pathogen. Escherichia. coli was isolated in blood cultures from one patient, and Klebsiella pneumoniae in another, giving significant positive blood cultures in 10% of patients. Significant pathogens were isolated from non-blood culture samples within 72hours of admission in 7 further patients. Of the 10 patients with significant pathogens isolated from admission samples, the pathogen was susceptible to piperacillin/tazobactam and co-amoxiclav for 9 patients (90%). The only non-susceptible isolate was a Meticillin resistant Staphylococcus aureus (MRSA) from a skin swab

Discussion: Concordance between paramedic diagnosis of sepsis and emergency department diagnosis appeared high, although further confirmation of accuracy is needed using objective sepsis criteria. Good documentation of compliance with the protocol was noted in the paramedic records. However, subsequent changes to the definitions and severity scoring of sepsis using the National Early Warning System (NEWS) parameters mean the protocol and associated patient group direction (PGD) require review and updating regarding criteria for inclusion.

Empirical coverage of subsequently identified pathogens in this group of patients was high, and would indicate adequate spectrum of cover could be maintained if the antibiotic used in the protocol was narrowed from piperacillin/tazobactam to co-amoxiclav. 

Blood culture contamination rates are high in this initial cohort and require further investigation.

065
Antimicrobial stewardship ward rounds using an app-based data collection tool – use in the Eastern Cape, South Africa

Abstract - 065

Poster 065

Antimicrobial stewardship ward rounds using an app-based data collection tool – use in the Eastern Cape, South Africa

Francis Collett-White1, Ya-Ying Wang2,3, Sarisha Singh2,3, Ncumisa Tafeni4, Andy Parrish4,5, David Stead4,2
1Thames Valley and Wessex Leadership Academy, Oxford. 2Frere Hospital, East London, South Africa. 3Rhodes University, Grahamstown, South Africa. 4Cecilia Makiwane Hospital, Mdantsane, South Africa. 5Walter Sisulu University, Mthatha, South Africa

Introduction: The challenge of increasing antimicrobial resistance and nosocomial infections are a significant issue in South Africa. This calls for effective interventions from front-line healthcare professionals to improve infection control and antimicrobial prescribing. We aimed to improve antimicrobial stewardship through multidisciplinary and pharmacy led ward round, utilising app-based audit and feedback.

Methods: Antimicrobial stewardship ward rounds were held monthly in 4 medical inpatient wards at Frere (FH) and Cecilia Makiwane Hospital (CMH), East London, Eastern Cape. Every patient was reviewed to assess the appropriateness of the antibiotic prescription by an Infectious Diseases consultant and the need for infection prevention control (IPC) measures (unnecessary IV cannula, urinary catheter, or isolation of an infective patient). Antibiotics for tuberculosis and prophylaxis were not included. An adapted AMS ward round for pharmacists was developed and used on surgical wards. Antibiotics were reviewed to assess compliance with clinical guidelines and the results were immediately fed back to ward doctors. A free iOS and android app, EpiCollect5, was used to develop a data collection tool that was used from February to July 2017 during AMS ward rounds. Data was stored securely online and downloaded to Microsoft Excel for descriptive data analysis.

Results: The prevalence of antibiotic use on medical wards was 19% (73/375) at FH and 26% (29/128) at CMH. Compliance to EML clinical guidelines and/or Infectious Diseases specialist opinion was 81% (59/73) at FH, and 74% (23/29) at CMH. The most common impact of the AMS ward round on the antibiotic prescription was stopping the antibiotic (26%, 19/73) at FH and changing the course duration (21%, 6/29) at CMH. An IPC intervention was required in 12% (45/375), and 6% (8/128) of patients at CMH and FH respectively.

Discussion: App-based data collection of AMS ward rounds allows for efficient auditing and objective feedback to clinicians. Antibiotic prescribing and compliance was higher at FH compared to CMH, and the need for IPC interventions was also higher. Clinical teams can find feedback motivational and this can be used to encourage their educational development. The AMS ward round app can be used by pharmacists in less well staffed departments to promote more prudent antibiotic use and guideline compliance. The use of app-based data collection for AMS ward rounds has continued for almost 18 months and an amended tool has been rolled out in George Hospital, Western Cape, South Africa.

066
Assessing and improving compliance with National Scottish Antimicrobial Group (SAPG) antibiotic prescribing quality indicators: a quality improvement project

Abstract - 066

Poster 066

Assessing and improving compliance with National Scottish Antimicrobial Group (SAPG) antibiotic prescribing quality indicators: a quality improvement project

Emilie Bellhouse
Hairmyres University Hospital, East Kilbride

Introduction: Antibiotic resistance is a significant public and global health risk. Resistance is directly linked to selective pressure of antibiotics and therefore every day antibiotic use. Longer durations of antibiotic treatment are associated with increasing resistance. There is evidence to suggest that early intravenous to oral switching (IVOST) reduces overall antibiotic days, without significantly affecting morbidity and mortality; a clearly documented antibiotic review within 72 hours facilitates early IVOST. Therefore, interventions to improve the rate of documented antibiotic review within three days may optimise antibiotic usage. This project aimed to assess compliance with National Scottish Antimicrobial Group (SAPG) antibiotic prescribing quality indicators, which align with the Public Health England Start Smart then Focus objectives.

Methods: Scottish Patient Safety Quality improvement methodology was used to assess baseline compliance with standards then “Plan Do Study Act” cycles were used to test interventions. Data were collected weekly from December 2017 for a 26 -week period on a respiratory ward in a district general hospital. Interventions focused on intravenous (IV) review with documented antibiotic plan. Interventions tested by the junior doctor included clarifying IV antibiotic plan on ward rounds, peer to peer informal education, discussing IV review documentation with consultants in receiving ward, and formal consultant feedback by e-mail.

SAPG standards state 80% of patient should have a review of their IV antibiotics within 72 hours with a documented antibiotic plan. Accepted antibiotic reviews include: IV to oral switch, Escalate, de-escalate, and continue with reason given. An analysis of the features included in the review documentation was also conducted on a subset of patients. Documentation of infection management plan was subdivided into: Patient information; Clinical detail; observations including NEWS (National early warning score) score and blood results; microbiology results; antibiotics review including day of antibiotics and current treatment regimen.

Results: Over the collection period 192 antibiotic episodes were audited and 119 IV antibiotic episodes were assessed. The median rate of documented IV review was 83% and the mean was 87%. Data before interventions showed 57% patients had a review that included a clearly documented antibiotic plan (Continue with no reason was deemed poor quality). After interventions the median increased to 84%. 

34 patients had infection management documentation audited. Patient descriptors were included in 25% patients, although allergy status documentation was 0%. Details of microbiology results (awaited or reported) were only present in 12% of reviews. The current clinical diagnosis was documented in 63% of patients and details of physical examination in 58%. 62% patients had NEWS or one or more observations included (Temperature, Respiratory rate, Pulse, Blood pressure, oxygen saturation). C-reactive protein (CRP) was recorded in 58% patients. Current antibiotic treatment was recorded in 48% of patients and day of antibiotic treatment in 12%.

Discussion: Good compliance with IV antibiotic review was seen and the current SAPG standard was met. Progress was made on improving the documentation of antibiotic plans and reducing the use of continuing antibiotics without a documented reason. Peer to peer discussion was highlighted as effective although asking for an antibiotic plan on the ward round was less effective as consultants readily discussed, but didn’t always document this at the ward rounds tested. It was challenging to influence the documentation of consultants. A concurrent project by nursing staff on antibiotic duration on the drugs kardex helped to raise the profile of antibiotic prescribing on the ward. We also found patient flow affected interventions aimed at the receiving team. The audit of documentation of infection management plan highlighted the variability of documentation in IV review. The data collected in the study will support SAPG in their work to create guidance on antibiotic review. 

067
Emphasis framing and antibiotic prescribing

Abstract - 067

Poster 067

Emphasis framing and antibiotic prescribing

Eustacia Hamilton1, Philippa Lilford2, Fergus Hamilton3
1Weston General Hospital, Weston Super Mare. 2Severn School of Psychiatry, Bristol. 3UHBristol

Introduction: Sepsis and antibiotic resistance (AR) are two great challenges in modern medicine. In some quarters, there is a perceived conflict between the two, as one pushes antibiotic prescribing, whereas the other pressures non-prescribing. In emphasis framing, focus is placed on a particular aspect in order to encourage that interpretation. 

For example, an electronic health record (EHR) system might provide information about antibiotic resistance automatically when prescribing antimicrobials, in order to ‘nudge’ appropriate prescribing.

We aimed to investigate if emphasis framing affects prescribing habits. We performed a randomised, blinded web based survey of doctors with fictional cases and simulated information about both sepsis and AR to see if this altered prescription habits.

Methods: We performed an online survey of clinicians, mostly hospital doctors. Participants were unknowingly randomised to receive information about either sepsis or antibiotic resistance (2 short paragraphs) prior to reviewing four clinical cases, all of which had a borderline need for antibiotics. Prescription habits, confidence in prescription, and time taken to answer questions was recorded, using a web based survey. Participants were unaware they had been randomised to wither arm.

We aimed to estimate differences in antibiotic prescribing due to the framing from the prior information

Results: 62 participants performed the study, with most providing complete data. Most were junior hospital doctors. In the sepsis group, they viewed the sepsis information for 16.9 seconds, while in the AR group this was 9 seconds (p<0.05, two tailed t-test). There was no difference in antibiotic prescribing, with both groups giving antibiotics commonly (83% sepsis, 81% AR, p=0.66). In a secondary analysis, including multivariable logistic regression with time spent on the survey and hospital grade included as a covariate, there was no significant association found.

Discussion: In this simulated web based survey, providing information about sepsis or AR made no difference to prescribing habits, despite being read for more than ten seconds. Cognitive biases have been shown to affect medical decisions but this varies depending on the level of risk and the type of heath decision involved. Antibiotic prescribing in clinically borderline cases does not seem to be influenced by emphasis framing. This might have clinical implications for EHR’s that aim to ‘nudge’ appropriate prescribing in infection.

Major Sponsors