**Standard disclaimer: these are my notes taken during the sessions, as accurate as I can make them, please let me know if any inaccuracies. I’m putting them here as others have said they found them useful. However worth re-checking anything before you re-quote it. If this continues to be useful, I’ll continue. If it’s clear i’m not reaching the standards of accuracy I’m aiming for, I’ll review future posting plans.**
Day 1 take home messages were
1) Whole genome sequencing is not just a research technique, it works, and is being used by a few groups who are pushing the field forward rapidly to revolutionise the study of transmission, and patient care. I may be biased coming from the Crook/Peto Group, but based on what was shown today, I think the case is very strong indeed.
2) To properly understand transmission requires understanding of phylogeny. To do phylogeny well requires rooting the tree with a proper range of samples, not just your local isolates. I’m going to quote Edward Feil because the quote was awesome (see later post) on the need for a shared central repository or database of isolates “You may say I’m a Dreamer, but I’m not the only one”.
3) Widespread multidrug resistance is not future, it’s now, and it’s progressing rapidly, year on year the figures are looking drastically worse. I remember ECCMID 2013 and the realisation that ESBLS were endemic in many parts of the world being shocking to me. It feels very much like the same conversations are being had about Carbapenem-Resistant Enterobacteriaceae that were being had about ESBLs 2 years ago.
In depth session descriptions:
1) WGS- What the heck?
Dag Harmsen started out, musing on the problem of backwards compatibility of all this new WGS data, and what we can do about it. He noted we have 15yrs of MLST data/similar (spa types/MIRUVNTR etc) and many more of pulsotypes an asked whether we use WGS data to give these, and thus compare with the past? He says literature says: yes to MLST, Spa, Salmonella serotyping, no to MIRU VNTR. Can we use WGS to predict pulsotype – nope. Tools available but don’t accurately predict gel picture: epigenetics and MGEs get in the way. Thus CDC are sequencing all their historic isolates from their pulsotype database to ensure forward compatibility.
He said what we need is available clinically useful information – enables visualisation of outbreak in 4D Gave a fantastic screen cap of Seqsphere user interface (proposed) which showed map data, phylotree and time graph, interactive (proposed). The case he made was, I think, that what we really need is a well curated catalog of validated relevant genetic markers which we have experience with, and which can be used for analysis LIKE MLST.
Matthew Holden was next, with the question – can WGS be used for staph/MRSA outbreak analysis? The answer was conclusively yes. Quoted an extensive array of papers from the Sanger/Cambridge/Wellcome group showing impressive results. Of note, coined the phrase ‘Index Dog’. Poor Mojo (Index dog). Healthcare worker spreading MRSA. He very much emphasised that in Staph. aureus there is significant within-host diversity (a ‘cloud’) which needs to be taken into account in any study design when sampling .Quoted a Paterson paper – due to cloud of within host diversity you could sequence 400 colonies and still not identify all variants. Clinically relevant remains to be determined.
Frederick Laurent from France gave a talk on ‘what would a clinician want and are we there yet’. His view – it depends! There’s promise, but only an option if you’re a teaching hospital with a genomics unit! Quoted the study from our group by Claire Gordon in 2014 on defining and validating use of NGS to predict Staph. aureus resistance phenotype, which I can’t recommend enough, mainly to demonstrate how much work is required to get it right enough for clinical use.
Next Prof. Edward Feil from the University of Bath, on WGS for identifying high risk clones. Useful definition of HiRICs as defined by the TROCAR FP7 group – clinically important arm, colonisation efficiency, highly transmissible, virulent and increased tenacity in the environment. He notes ESBLs CREs the most clinically significant. Then says he isn’t going to talk about them, but about Staph! Honesty! He notes the difficulty of inferring transmission/relatedness with only local samples – to properly base phylotrees, need representative sample from other places. He then had a Lennon moment, talking about what he would like to see – all centres contribute a selection of representative isolates or sequences to a database which you can draw all phylotrees with. “You may say I’m a Dreamer, but I’m not the only one”. He also gave a demonstration (screen cap) of www.microreact.org (which is now down) which looked just like the seqsphere visualisation, in a wonderful piece of convergent evolution.
Antibiotic Stewardship session:
Take home message: lots of people are doing lots of very worthy things. They probably work a bit. The problem is massively different depending on where you are.
The PRIOAM stewardship intervention was presented (Cinneros Herreros) which involved face to face counselling with an ID specialist with leadership role and based on a case the learner had treated and structured questionnaire. GREAT! Saw 1140 to 791DDDs/100 admissions sustained decrease, with increase in blood cultures taken and positive. Made a very good attempt to demonstrate change in resistance but difficult to demonstrate effect due to other interventions and possible changing epidemiology. Decrease in acinetobacter. Interesting as incorporated face to face and successful and decent design of analysis. A huge step forward from the before and after audits of the past!
Presentation of NAPS (national survelliance/reporting of abx use in Canada) – Chen showed a really nice user interface which allows you to input data and also to benchmark your use against other comparator hospitals. Also allowed them to identify particular cases where prescribing quality was poor and target these rather than blanket approach (surgical prophylaxis, resp tract infections and 3rd gen cef prescribing).
Esmita Charani – effect of smartphone app. Take home message – if your compliance is already good due to high presence of good team, not much effect. Also released around start smart and focus, and likely this had big effect in improving documentation of stop date rather than app. However again, an excellent study which actually measured its outcomes well. She noted that app is designed and used for decision support and juniors liked this.
A.Chow gave a demonstration of Computerised Decision support system in Singapore – great design of tool , ok study If you prescribe (computerised) normal abx, voluntary use of CDSS. If restricted abx- mandatory. Brings up a screen of demographics, clinical info .micro info, recommendations and then accept/override. Really nice. 83% compliance, medics and on call Doctors more compliant, less compliant – Sepsis (my patient is ill goddamit!) and uti treatment. I want it for my work!
C Kwok – Acceptance of stewardship recs triggered by 48hr abx -515 pts reviewed, generally well accepted recommendations, Interestingly high >70% compliance with stop. 21% compliance with ‘get a formal ID consult’ . ?cost?
So by the time I was at this my brain was a bit kaput… It was basically lots of reports of people study resistance profiles by choosing different sampling criteria (bug/strain/gene/phenotype) and then defining their population using phenotyping and PCR, sometimes a bit of sequencing. Huge mix in results, take home message is that carbapenemases are everywhere and being talked about like they’re normal. Lots of stuff on MGEs. Bad stuff abounds.
Ny – ESBL prevalence in sweden. To try and get representative sample sent 11400 questionnaires to random people in sweden, got 2130 faecal samples. Many Swedes seem quite happy to send their poo for science! Found 4.7% prevalence of ESBLS (101 positive) all E.Coli! Bit odd- picked up on question from audience – Teresa Coque on the panel asked about their methods – she said Maconkey agar/MALDI – panel noted you often need Citrate to pick up Klebs. Anyway – travel and non-pork diet was a risk factor for ESBL carriage (OR 0.5!) likely indicator of something else? Mainly CTX-M-15 IncF H30RX. I’m ignoring proper science and starting the Campaign for Eating Bacon to Combat Resistance.
Jones-Dias- widepread detection of ESBLs in agriculture in Portugal, with Intensive farms having more resistant bugs and Organic farms less. Most isolates Pseudomonas and Acinetobacter but fair number of E Coli and Klebs.
Castanheira – prevalence of amino glycoside resistance genes. This was just a mess (the reality, not her talk). 200 isolates chosen from bank based on differing amioglycoside resistance pattern (AMI GENT TOBRA). Basically loads of different genes and variants often co-carried /mixed giving very different resistance patterns. Admirable study. Says in conclusion what we really need is curation of genes and nomenclature! Based on this I would highly highly concur.
Teo – molecular characterisation of Carbapenem-resistant isolates in Singapore.Diverse range of carbapenemases!!! Oxa48, kpc, ndm1 – this ain’t clonal, this is real as it gets -multiple imports and spread.
Ghebremedhin – presented some data on CTXM isolates in Kenya suggesting that CTXM15 ST131 was the strain associated most with virulence genes and biofilm formation genes.
Castanheira’s presented data from the Turkish SENTRY program. Noted that Turkey not part of EARSS Estimates of Klebsiella CRE 3% previously – this study makes it 11% of Klebsiella pneumo. hospital isolates mero resistant, 16.5% imipenem resistant.
In conclusion – resistance, resistance everywhere.