- Volume 2, Issue 2, 2020

The Scottish Antimicrobial Prescribing Group (SAPG) is supporting two hospitals in Ghana via a Fleming Fund healthcare partnership to develop antimicrobial stewardship. Initial intelligence gathering suggests that antibiotic prophylaxis to prevent surgical site infection (SSI) is suboptimal. To inform a quality improvement programme we have reviewed the evidence for use of surgical prophylaxis in LMICs including staff behaviours and attitudes.

MEDLINE, Embase, Cochrane, CINHAL and Google Scholar were searched from inception to 22 July 2019 for trials, audits, guidelines and systematic review in English. Grey literature, websites and reference lists of included studies were searched. The following data were extracted; study characteristics, interventions, outcomes and recommendations. In view of heterogeneity between studies descriptive analysis was conducted.

Of 185 records screened, 26 studies related to SSI and timing of antibiotic prophylaxis in LMICs were included. The incidence of SSI is significantly higher in LMICs compared with high income countries, recording of infection surveillance data is poor and a lack of local guidelines for antibiotic prophylaxis. Several projects in Africa have reported reduction in SSI with single dose preoperative antibiotic use compared with post-operative prophylaxis and a reduction in cost and nurse time. Despite evidence to the contrary, many surgeons continue to use post-operative antibiotic prophylaxis.

Education to improve incidence of SSI in LMICs through appropriate antibiotic prophylaxis can be successful. Interventions must include local context and address strongly held beliefs. The establishment of local multidisciplinary teams will promote ownership and sustainability of change.

A facilitated learning session was delivered and all clinical and non-clinical GP Practice staff, were invited to attend. Evidence supporting best practice in managing UTI, local prescribing data and local practice was considered and discussed. The session concluded with action planning to reflect on current practice and identify areas for improvement.

National data held by NHS National Services Scotland was used to evaluate prescribing across all NHS board regions in Scotland focusing on total use of antibiotics and use of trimethoprim and nitrofurantoin which are used solely for UTI.

Analysis of national prescribing data comparing the implementation region with the rest of Scotland suggests the intervention has been successful. Comparing data from prior to and for one year after the intervention period reductions for the intervention board versus the rest of Scotland were as follows: all antibiotics 7.29% vs 6.11%; trimethoprim 6.14% vs 3.73%; nitrofurantoin 3.92% vs 0.56%.

Quantitative evaluation provides evidence of the impact of ScRAP on prescribing rates. This suggests that improved practice has led to reduction in unnecessary use of antibiotics for acute and recurrent UTI.

A facilitated learning session was delivered, usually by a Prescribing Support pharmacist. All GP Practice staff, both clinical and non-clinical, were invited to attend to promote a team approach to change. Local practice was considered using process mapping and each GP practice completed an action plan with changes they intended to implement. Facilitators and participants were encouraged to complete an on-line feedback survey about the learning session. Action plans from 200 GP Practices in one health board region (21% of Scottish population) were analysed using NVivo 12 software to identify key themes.

Completed surveys from facilitators and participants showed both groups were positive about the content, approach and length of the session. Action planning was welcomed as a way to embed learning in practice. Some minor changes to content were suggested. Changes in practice included patient education to promote self-management and a reduction in the number of urine specimens sent to microbiology laboratories. A whole team approach to management of patients with suspected UTI increased the success and impact of the changes made.

Evaluation of action plans will be helpful for other GP Practice teams seeking to make improvements. Feedback from facilitators and participants will be used to inform update of the programme content.

As part of a move to improve antimicrobial stewardship NHS GGC recently implemented identification to species level and Vitek 2 sensitivity testing on all urinary isolates. As part of this we have noticed an increased number of isolates of the Streptococcus bovis group in urinary samples. It is possible that these were previously labelled as Enterococcus spp. The significance of S. bovis isolates in urinary tract is uncertain.

Methods

We reviewed a year of S. bovis urinary isolates to ascertain whether there is any association with GI malignancy or endocarditis as is recognised with blood culture isolates of this group.

54 isolates were reviewed dating between Feb 2018 and 2019 allowing at least 6 months follow up. 94% of cases were in females. 46% of isolates were mixed usually with a member of the Enterobactereaceales. There were no cases of either known or newly diagnosed endocarditis amongst this patient cohort. 9% of cases had a known GI malignancy.

9% of our urinary isolates of S. bovis were associated with a known GI malignancy. This is compared with 11% of blood culture isolates also from our unit in 2016.

There is some evidence of Pseudomonas aeruginos across-infection between patients with non-cystic fibrosis bronchiectasis (NCFB), and clear evidence in Cystic Fibrosis. Haemophilus influenzae (H. influenzae) is the more common pathogen in NFCB patients, yet cross-infectionremains unexplored. We present the novel application of culture-independent Multilocus Sequence Typing (MLST) to screen for cross-infection of H. influenzae in NCFB in both culture-positive and -negative samples.

Methods

We interrogated DNA from 32 sputum samples (26 patients) in our NCFB biorepository, who were known to have H.influenzae in their sputum by preceding 16S rRNA sequencing. Fragments of 7 H. influenzae housekeeping genes were amplified and sequenced. Sequence types were allocated via the MLST scheme. For 5 patients, multiple sputum samples taken at least 4 months apart were assessed longitudinally.

Culture-independent MLST identified 31 of 32 sputum samples as harboring H. influenzae. Of these, 26 were positive for H. influenzae using culture methods. 25 of the 26 culture-positive samples were MLST positive. All 6 culture-negative samples were MLST positive. A MLST sequence type (ST) was allocated to 27 of 32 sputum samples. Five patients had multiple sputum samples with matching STs, indicating strain stability and the consistency of MLST. Two patients who were known household contacts had matching STs and possibly transmitted H.influenzae in their household. The remaining 15 STs were unique, suggesting no evidence of cross-infection.

Culture-independent MLST identifies H. influenzae in culture-negative patients with NCFB and is a potential screening tool for cross-infection. This study did not reveal potential cross-infection events in this cohort.

To address increasing antibiotic use in acute hospitals, the Scottish Antimicrobial Prescribing Group developed a quality improvement (QI) initiative to support reliable review of patients started on intravenous (IV) antibiotics within 72 hours. This will reduce unnecessary continuation of antibiotics, ensure personalised treatment and appropriate IV to oral switch with associated benefits for patients of reduced risk of device related infections and potential for earlier discharge from hospital.

Methods

Using data obtained from the Hospital Medicines Utilisation Database, a national database of secondary care medicines use in Scotland, we examined trends in IV antibiotic use between 2013 and 2017. We then projected the current trend forward to 2021 to inform development of national indicator to optimise IV antibiotic use.

In 2017, IV antibiotics accounted for 32.9% of all antibiotic use in Scottish hospitals. Annual IV antibiotic use (defined daily doses per 1000 population per day) increased by 20.5% between 2013 and 2017. We estimated a further projected increase of 12.5% between 2018 and 2021. To measure the impact of our QI initiative SAPG agreed to employ a national indicator with a target that ‘use of IV antibiotics in hospitals will be no higher in 2021 than it was in 2018’.

This national indicator will evaluate progress with achieving reliable and timely review of IV antibiotic therapy to reduce hospital antibiotic use and contribute to reduction in total antibiotic use in humans which is a key ambition of the UK AMR National Action Plan.

We undertook a retrospective observational study at a central London teaching hospital to investigate antimicrobial prescribing in adult patients undergoing appendectomy between Jan2019 and Jun2019.

Results.

A total of 173 patients [median age 31year] were analysed; 7 and 167 had undergone an elective and urgent appendectomy respectively. 163/173 [94.2%] received antimicrobials peri-operatively. Compliance with local guidelines was 40.5%; a lack of aminoglycoside in combination with beta-lactam was common (58%). 126/173(72.8%) patients received antimicrobials on discharge, median 7.0 days total antimicrobials. Readmission rates within 30 days of surgery were 6.4% but unrelated to antimicrobial prescribing (p=0.8). Enterobacteriaceae was the most commonly identified pathogen (n=45), with high co-amoxiclav resistance reported (31%).

Conclusions

Combination co-amoxiclav plus aminoglycoside is advised peri-operatively for appendectomies yet adherence is poor. Despite local co-amoxiclav resistance, aminoglycosides are often omitted. The number of patients treated with post-op antimicrobials and the duration of therapy is greater than comparable published studies. Lack of standard definitions for complicated appendectomies makes it difficult to identify patients that do benefit from post-operative antimicrobials thus overprescribing occurs. Agreement on the optimum duration of treatment also is unclear and results in likely excessive prescribing.

We undertook a retrospective observational study at a London teaching hospital to investigate the incidence of invasive GBS peri-partum in both mother and child between 1/4/16 and 31/3/19.

There were 16,869 live births recorded at Chelsea & Westminster Hospital during the study period. Operative, caesarean and spontaneous delivery accounted for 16.7%, 34.9% and 48.2%; with incidence of neonatal GBS bacteraemia being 0.18%, 0.08% and 0.11%, respectively. 20 (0.12%) neonates had invasive GBS infections, with 7/20 neonates having concurrent maternal GBS bacteraemia. 34(0.22%) ladies had GBS bacteraemia peri-partum; none had GBS isolated prior to delivery. 543 women had a positive GBS clinical isolate prior to labour;69.1% received GBS prophylaxis peri-partum with no invasive GBS transmission. Despite no prophylaxis in 168/543 GBS colonised women, no invasive GBS neonatal cases were identified.

Conclusions

Our local data identifies a low prevalence (0.12%) of invasive GBS infection in newborns. Maternal GBS bacteraemia is more frequently observed and associated with concurrent newborn GBS bacteraemia. Antimicrobial prophylaxis adherence peripartum although recommended is suboptimal, even though no invasive GBS transmission was identified. All invasive GBS infections were not known to be colonised pre-delivery. The current recommendations for targeting recent GBS culture do not appear to be a sensitive predictor of invasive peri-partum GBS infection.

Spontaneous bacterial peritonitis is an acute infection of ascitic fluid that is not related to an underlying intra-abdominal pathology.

This is a retrospective study of data on ascitic fluid samples taken during the period of January to December 2017 to assess the management of SBP patients and associated mortality. Kaplan-Meir method was used to estimate survival probability at 30 and 90 days. Cox proportional models were used to evaluate aetiology and causative organism in predicting mortality. Analyses were done using R: A language and environment for statistical computing.

Total of 53 patients were identified as having SBP based on ascitic fluid cultures. Sixteen of these also fulfilled the national criteria of ascitic neutrophil count of more than 250 or total cell count of more than 500.

Thirty-three (62.8%) samples grew pathogenic organisms. The most common pathogenic organisms were E coli (30.3%) followed by Enterococcus (21.2%) – in which four were ESBL tagged – Klebsiella (6%) and Staph Aureus (6%). In majority of the patients, the underlying aetiology of ascites was secondary to ALD (68.9%), Malignancy (11.8%) and NASH cirrhosis (11.8%). In this cohort, the overall 30-day survival was 64.2% and 90-day survival was 43.4%. About half mortality occurred during the same admission to hospital.

E coli was the most commonly isolated organism. Alcoholic Liver Disease was the most common underlying aetiology in SBP patients. SBP is associated with a high 30-day and 90-day mortality. Isolation of pathogenic organisms was associated with a four-fold higher mortality than non-pathogenic organisms.

Standard of care for management of Staphylococcus aureus bacteraemia (SAB) is 2-4 weeks intravenous (IV) flucloxacillin or glycopeptide. Ceftriaxone (CRO) is used to facilitate management of SAB under out-patient antimicrobial therapy (OPAT) services once patients are medically stable, however published data on this approach are limited.

Retrospective review of SAB cases at Homerton Hospital: 1st August 2015 to 31st July 2018. Cases were identified from the microbiology database and clinical data retrospectively collected from electronic patient records.

83 cases of SAB were included. Median age was 56 years (IQR 45-74); 53 (63.9%) were male. 70 (84.3%) had complicated SAB, 4 (4.8%) had MRSA bacteraemia and 11/80 (13.8%) were PVL positive. After excluding patients who died or were transferred whilst on IV therapy; 8/11 (72.7%) uncomplicated SAB patients and 29/55 (52.7%) complicated SAB patients received the standard duration of IV anti-staphylococcal therapy. Median length of stay (LOS) was 32 days (IQR 16-52.5). 30-day mortality was 9.6%; in hospital mortality was 14.5%.

Eight (8/83, 9.6%) patients switched to CRO prior to completion of standard IV flucloxacillin therapy to facilitate OPAT. Median length of IV flucloxacillin in this group was 12 days (IQR 7-16). Ceftriaxone MIC was performed on 1/8 isolates (3mg/L). 7/8 had complicated SAB. Median LOS was 13 days (IQR 9-17). There were no deaths or relapsed infections. 1 patient developed C. difficileinfection on CRO.

In this cohort ceftriaxone was a safe and effective follow-on therapy from flucloxacillin for management of SAB and allowed reduced LOS.

Only five cases of Salmonella paratyphi infective endocarditis (IE) have been previously reported. Two were paediatric patients, one other had an underlying cardiac lesion, and all were living in endemic areas. Here we present the first case of salmonella paratyphi IE in a returning traveller.

A 61 year old Indian born British national with no significant past medical history or recent dental procedures presented with a 4 week history of fever and weight loss following travel to India. Whilst abroad she had a self-limiting episode of diarrhoea with fever for 1 week with no antimicrobials. Three weeks after her return the fever returned (without diarrhoea) and she presented to hospital 4 weeks later.

Examinations findings were; haemodynamically stable, temperature 39.5°C, no splinter haemorrhages, no lymphadenopathy, soft systolic murmur best heard at parasternal edge.

Investigations revealed raised inflammatory markers (CRP 80) and blood cultures grew gram negative bacilli within 24 hours which cultured as Salmonella paratyphi A.

Trans-thoracic Echocardiogram showed echo-bright thickening of the non-coronary cusp of the aortic valve, which was confirmed as a small vegetation (0.5cm x0.2cm) on trans-oesophageal echocardiogram.

IV antibiotics were given for a total of 6 weeks and repeated TOE at 1 month showed a healing vegetation. Repeated imaging is awaited.

This interesting case in an otherwise healthy patient highlights the need to be vigilant for unusual organisms causing infective endocarditis – a relatively common condition - in returning travellers.

292 patients (174 male, 118 female) median age 54 (range 17– 97) April 2017-June 2019. 148 patients with SSTI and 143 DSI. 140 patients previously treated as inpatients. 141 admissions avoided; 3540 bed days saved; 267 patients treated with H-OPAT; 21 C-OPAT; 4 S-OPAT. High cure rate of 100% in SSTI and 97% in DSI (3 re-admissions, 1 self-discharge). Low complication rate with no death, MSSA/other bacteraemia or CDT-associated diarrhoea. Low complication rate of 0.7%: rash in 1 patient (on ceftriaxone) and line blockage in 1 patient.

The aim of the OHPAT service is equality of effectiveness and safety compared to inpatient care. This has been achieved in the ongoing low rate of complications and adverse outcomes. The service treated a wide range of DSI which contribute a larger proportion of bed days for much fewer cases than SSTI. Significant numbers of hospital stays were avoided or shortened. OHPAT has been able provide complex regimes with home delivered and self-administered dosing.

This prospective study was established to develop and evaluate methods to harness pathogen sequencing in the clinical microbiology environment. We describe new insights into the clinical benefits of using whole-genome sequencing (WGS) for outbreak investigation with solutions for the practical barriers to implementation in clinical settings.

Methods

Surveillance software (ICNet) and statistical process control charts (SPCs) detected potential outbreaks of meticillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), carbapenem-resistant Enterobacteriaceae (CRE), Enterobacteriaceae producing extended spectrum β-lactamases (ESBLs), Listeriaspp. and optrAgene positive enterococci. Isolates were sent to reference laboratories for conventional typing and the Infection Group, School of Medicine, University of St Andrews for WGS (Illumina Inc, San Diego, CA, USA) and bioinformatic data analysis.

Over 400 isolates have been sequenced to date. WGS replaced multiple typing techniques in a Pseudomonas aeruginosa ICU outbreak. It confirmed two patient’s Listeria spp. isolates were indistinguishable prompting hospital kitchen inspections, identified a patient to be carrying two strains of VRE and confirmed that vancomycin-sensitive Enterococcus faecium isolates related to a VRE cluster. We observed five main barriers to implementing WGS (infrastructure, performance/quality assessment of data, isolate selection, clinical result interpretation and database management).

WGS is beneficial in outbreaks of uncommon organisms and when conventional typing cannot show whether isolates are linked or not. Identifying barriers assisted us in developing a clinical decision aid that can be used by clinicians when applying WGS to outbreak investigations.