The SNAP Trial with Professors Tong and Davis

In this episode of "In the Interim…", Dr. Scott Berry interviews Professors Steven Tong and Josh Davis about the SNAP platform trial for Staphylococcus aureus bacteremia. The discussion covers SNAP’s rationale, large-scale adaptive design, methodology, and operational execution at approximately 150 hospitals in 13 countries. Key statistical questions, domain results, pediatric-adult analysis, and global implementation strategy are explored in depth. Listeners will find clear examples of how adaptive platform trials can efficiently address clinically relevant questions in infectious disease, while highlighting the nuances of trial design, statistical thresholds, and network collaboration.

Key Highlights

• High and unchanging mortality for Staphylococcus aureus bacteremia—over one million deaths annually.
• SNAP leverages silo-based structure (MSSA, MRSA, PSSA) and factorial domains for simultaneous, efficient investigation of treatments.
• Cefazolin shown non-inferior to flucloxacillin for MSSA with lower related acute kidney injury.
• In PSSA, penicillin demonstrated significantly less toxicity and favorable mortality signal over flucloxacillin; mortality difference did not meet the statistical superiority threshold.
• Futility reached in the adjunctive clindamycin domain for effect on 90-day mortality.
• Both adults and children enrolled, with pediatric results using statistical borrowing from adults in line with FDA Bayesian guidance.
• Ongoing platform expansion includes bacteriophage therapy, antiplatelet domains, and evaluation of diagnostic strategies.
• Statistical leadership: Dr. Anna McGlothlin (Berry Consultants), Dr. Julie Marsh (statistics lead).

For more, visit us at https://www.berryconsultants.com/

Creators and Guests

Host

Scott Berry

President and a Senior Statistical Scientist at Berry Consultants, LLC

What is In the Interim...?

A podcast on statistical science and clinical trials.

Explore the intricacies of Bayesian statistics and adaptive clinical trials. Uncover methods that push beyond conventional paradigms, ushering in data-driven insights that enhance trial outcomes while ensuring safety and efficacy. Join us as we dive into complex medical challenges and regulatory landscapes, offering innovative solutions tailored for pharma pioneers. Featuring expertise from industry leaders, each episode is crafted to provide clarity, foster debate, and challenge mainstream perspectives, ensuring you remain at the forefront of clinical trial excellence.

Judith: Welcome to Berry's In the
Interim podcast, where we explore the

cutting edge of innovative clinical
trial design for the pharmaceutical and

medical industries, and so much more.

Let's dive in.

Scott Berry: Uh, welcome everybody
back to, in the interim, I'm your host,

Scott Berry and I'm joined by a couple
investigators from down under, uh, we are

gonna talk about the SNAP platform trial.

So let me first introduce
our guest for today.

I have Professor Steven Tong and
he's a practicing infectious disease

physician at Royal Melbourne Hospital
and a professor in of infectious

diseases and clinical trialists.

At the Doherty Institute, the
University of Melbourne, and one of

my more favorite titles, he's the
head of Adaptive Platform trials.

Uh, that's really cool.

At Melbourne, uh, center
for Clinical Trials.

He co-leads the Streptococcus Aus
Network Adaptive Platform Trial.

And I hope that's the last
time I have to say that.

Uh, and which is recruiting participants,
150 hospitals, 13 countries, and we're

to get into the SNAP trial, uh, as this
is the topic of today's discussion.

And he is joined by Professor Josh
Davis, who's also an infectious disease

and general internal medicine physician
at the John Hunter Hospital in New

Castle, new South Wales, Australia.

He's also a professor of infectious
diseases at the University of

New Castle and Director of the
Infection Research Program.

At the Hunter Medical Research
Institute and he co-leads the

SNAP trial, uh, with Professor
Tong as well as the roadmap trial.

We probably won't talk, talk about
roadmap, but this is an adaptive

platform trial in management
of, of hip and knee infections.

Uh, and interestingly, Josh and
Steve did their PhD studies together.

Uh, in Darwin at the Menzies School of
Health Research in Darwin, and have had

a productive, ongoing relationship that
continues to grow for the last 15 years.

So, uh, uh, Josh and Steven,
welcome to, in the Interim.

Josh Davis: Thanks for

having

Steven Tong: Scott.

Yeah, it's, um, I enjoyed
listening to in the interim, so

it's a great honor to actually be
invited to come and talk on it.

Scott Berry: Wonderful.

It's, it's, it's exciting.

So first I, I wanna ask about Darwin.

Darwin is, uh, such an interesting place.

Uh, I've never been there.

Um, and, but it, it strikes me as a very
isolated area, kind of a unique area.

Seems like a really neat area to
do work in infectious diseases.

Is, is your work and and time
in Darwin part of this story?

Josh Davis: Yeah.

Um, it is, I mean, that's
where we met each other and

started collaborating, but also.

Where we, um, planned and ran the, the
Canberra one trial, which was at the first

trial we did together in MRSA bacteremia,
staph Aureus bloodstream infection.

That then led on to a, a, a series of
steps that ended up with the SNAP trial.

Um.

We were both there because it's a place
that I infectious diseases trainees,

um, often go to, to spend a few years
in Australia because it's fantastic

infectious diseases training environment
and research environment as well up there.

Steven Tong: So it's part of the tropics,
and so there's lots of tropical diseases.

There's a environmental
organism called myosis.

Which you only see up in Northern
Australia, at least in Australia,

there's lots of dengue malaria that
comes through from Southeast Asia.

So really great place for
infectious diseases training and

also, uh, first Nations health
is super important there as well.

So there's large Aboriginal population,
so a lot of us go there to learn

about, I guess, the disparities
that we see in our country.

Scott Berry: Mm-hmm.

Now I we're, we're gonna get to the SNAP
trial, but do you have a site in Darwin?

Steven Tong: We do, we do.

Absolutely.

So, and as Josh said,
that's where it all started.

We started studying staph aureus
because we saw high burden of

disease in the indigenous population
in the Northern Territory.

In fact, my PhD was done up in Darwin,
uh, specifically on staff ORs in

the northern, in Northern Australia.

Scott Berry: Hmm.

Okay.

So that, that comes to
the, the SNAP trial.

Uh, and the SNAP trial.

It says The aims are to improve treatment
outcomes for patients with Staphylococcus

Aureus bloodstream infections.

So tell me how this trial came about.

Steven Tong: Okay.

Uh, so staph aureus is a major
burden of disease globally.

Of the 33 top bacterial pathogens that
cause disease, it's actually number one.

So outside of tuberculosis of
the top 33, it's number one.

It's only one that's estimated to
cause more than a million deaths

per year, of which about 300,000,
are due to bloodstream infections.

So this is the most serious form of Staph
aureus infections, Staph aureus causes you

know, skin infections, uh, infections of.

Uh, other parts of the body, but once
it gets into the bloodstream, that's

when it's most serious and causes a
mortality rate of 20-25 per episode.

So it's a big deal.

Uh, and so it's important to
study as infectious diseases.

Clinicians, we see lots of this,
so every week in our hospital

we're on clinical service.

We are seeing, you know, two or three
patients per week with this condition,

and so it's one that really needs.

Study to try and improve outcomes.

That mortality rate of 25% or so per
episode has been stuck at that for

probably the last 30-40 years, I think,
and we haven't seen improvements in

our therapeutics for Staph bacterium

Scott Berry: So what, um, I, so
you're talking to a statistician

here, but, but, but it may help.

What is, what does it mean to have a
bloodstream infection and I, I suppose

it's, it's, it's moved from an organ
and the blood itself is infected.

What does that mean,

Josh?

Josh Davis: Yeah, well, um, I guess
technically it means we have found it.

We've detected it in the blood.

Uh, using blood cultures, but
generally it starts from the skin,

um, from either a skin infection
or through a cut or a scratch on

the skin gets into the bloodstream,
and it being in the bloodstream by

itself isn't the primary problem.

It's that the bloodstream is.

The root of transport to
everywhere in the body.

Um, and then so the staph aureus
germ can then spread to other organs

and often does do so, and it ends up
infecting bone, heart, liver, et cetera.

Um, yeah.

So bloodstream infection means the,
the bacteria rather than just sitting

on the skin has invaded into the body.

Scott Berry: Okay, so I could
have an infection, but it is.

Uh, it's, it's on the skin.

It's outside.

It has not gotten inside, uh, in it.

Okay.

So I, it, it, it's quite the impact
and just, just this, this is a

US thing that people in the US
are going to have this infection.

This is a first world problem.

This isn't just a Darwin problem.

Them,

Steven Tong: It's everywhere.

It's, it's, it's, it's very prevalent
in the US It's very prevalent in

probably less well described in
low middle income country settings,

but I'm sure it's prevalent there.

It just gets, there's so many other
causes of infections and, uh, burden

of disease that it, it doesn't quite.

You know, have the prominence that we
sometimes see in high income countries.

Uh, it's, it's an organism.

It is a really interesting organism
in that once it does get in

the bloodstream, it has a whole
variety of virulence factors.

That means it can escape
the immune system.

Uh, it can form what we call biofilms.

So if there's any foreign bodies
like pacemakers, um, heart valves,

prosthetic heart valves, prosthetic
joints, that it can, it gets really

sticky and it sticks to those.

And so it becomes very difficult to treat.

Uh, and it's not just antibodies,
therefore you need to remove

infected devices as well.

And in high income countries, as more
and more people get joints, you know,

replaced, get prosthetic valves, it
becomes more and more of a problem.

Scott Berry: Okay, so, uh, clearly
a, a high unmet need, uh, a million.

Uh, deaths a year, you said, or
case a million deaths a year and

things have not moved in in decades.

So what, what was the genesis?

Uh, uh, I mean, presumably
the trial was to change that.

Um, it, the, the genesis of then
the SNAP trial to investigate this.

Steven Tong: So we can follow on
from that story that Josh started.

He and I were seeing a lot of
patients with MRSA, so the more

resistant form, methicillin resistant
staph bloodstream infections.

In Darwin and we actually just
sat at a table and said, what

are we going to do about this?

And so we started a small pilot trial with
some, uh, colleagues in Australia that

got to a 60 patient randomized clinical
trial that proved some feasibility for

a larger study of about 350 patients
with MRSA bloodstream infections.

That one was called camera two, and
now we were involving four countries.

And once we finished that trial.

Uh, we started talking to Steve Webb,
who many of your listeners might know.

Uh, he, uh, heads up
Reman Cap in Australia.

Um, and Steve started talking to us
about adaptive platform trials and we

just thought, well, we've got these two
things coming together, staff bacteremia.

We've got some runs on the board
with, with a couple of studies now.

Uh, we'd like to apply an adaptive
platform trial to this particular disease,

and that's really where it started.

So talk those discussions with
Steve and Josh and myself.

We wrote a couple grants and the second
time we wrote that grant, it got up

and we therefore started the trial.

Scott Berry: Hmm.

Josh Davis: I guess maybe to add to that,
that that MRSA camera trial was about

combination antibiotic therapy, so using
two antibiotics instead of the standard

one to treat MRSA bloodstream infection.

And the next step, um, for that would be
that we wanted to do was to do a larger

trial of much many more patients of
the same intervention, but in parallel.

There are a whole lot of other questions
in management of this infection we

were interested in and that there had
been preliminary data develop, um,

either generated by us or other groups
that we wanted to turn into trials.

So, um, we could have done that, uh, in
serial, which would've taken 50 years,

but we could have done it all at the
same time, which is kind of what Snap is.

Scott Berry: Hmm.

Yeah, so we'll get to that.

This is a very cool factorial
trial where you're doing multiple

investigations on, on the same patients.

What I think it seems like part of this
story though, is the, the culture of, of.

Staph or bacterial infection research.

Is there a deep history of
clinical trials in this area?

You know what?

What's the background in this area?

Steven Tong: Yeah, that's
a great question, Scott.

And we, um.

I think in infectious diseases
are somewhat well behind lots

of other craft specialties.

So you know, we look at
the cardiology trials with

thousands of patients per trial.

They've been running trials for
myocardial infarction, I don't know,

since the sixties and seventies,
you know, worked out that we should

be giving aspirin to patients.

So really important questions, but we
just have not had a great culture, I

think, in infectious diseases of doing.

Similar kinds of trials, and you could
argue that staph aureus bloodstream

infections is our myocardial infarction.

And so we've been behind the
8 ball on doing these studies.

So as an example, with staph
aureus bloodstream infections,

there's probably been maybe 15,
20 trials over the last 20 years.

Uh, and the largest of
those is 750 patients.

And most of those trials sit
around the 200 patient mark.

And so there's just not been
large well powered trials and no,

none of those trials have found a
superior treatment to the control.

So some of them go non-inferiority trials.

Some have been superiority trials,
but not one of those has actually

found a treatment that's actually
improved patient outcomes.

So we've come into this space with our
CAMERA1 and CAMERA2 trials, and then

snap, uh, there was an Uh, editorial in
the Lancet written by some very prominent

colleagues, Vance Val and Tom Holland
in 2017 that really made a, a call to

say that our community needs to study
this disease better and priorities,

prioritize the right questions and form
clinical networks that can get sufficient

sample size to answer these questions
with sufficient power and precision.

Scott Berry: Hmm.

And hence, hence the SNAP trial.

Uh, I, I was able to find this, uh,
uh, new England Journal of Medicine

Group, uh, post made by Paul Sachs.

And at the end of the year he said,
uh, you know what, what, what should

we be happy for in infectious disease?

And number three on the
list was the SNAP trial.

Uh, in its study of, of staff, uh,
and he says things and, uh, some

things about the US that the largest
randomized study ever done for staff.

Uh, SAB finally brought clarity
to treatment questions that have

lingered since my fellowship days.

His fellowship days conducted across
multiple countries, uh, notably

excluding the us, which is he says,
a longer story and a longer rant.

Um, within this, um, and pointing out this
is being done in Australia and he talks

about the various results we'll get to.

So it seems that the community is
quite, uh, pleased and happy and

excited about what Snap's doing.

At the same time, I.

Josh Davis: Yeah, I mean, I, ID physicians
and, and microbiologists and people that

work in this area love this trial, but it
is, it's sometimes easy for us to forget

that not everyone in the world is an ID
physician and is that interested in it.

sometimes if, if people, you know, family,
friends who are not doctors, ask me about

what this trial's about, they're like, oh.

That doesn't sound very interesting, but

Id people love it.

Um, and we think it is really important
of course for, uh, given that this is

such a common illness and common disease.

So yeah, the reception from, from that
part of the, from ID people and those that

are interested in research and clinical
treatment in this area has been fantastic.

And in fact, it, it's what's grown
the network so quickly in terms of the

sites and countries coming on board.

Steven Tong: And I think
the success to date.

It does relate to the engagement of
all our colleagues around the world.

So people only recruit to this
trial because the questions are

clinically relevant to them.

As SACS is saying, these are questions
that have been around for 30, 40

years and we've just not organized
ourselves sufficiently to answer

them in the way that we should.

Scott Berry: Okay, so let's talk
a little bit about the trial set

up, the structure, the, the usual
clinical trial components of this.

So your patients come in with, uh,
staph infection, blood infection.

Do they have to have a resistance?

To get in the trial.

So a, a lot of this you talk about
P-S-S-A-M-S-S-A, um, what, what is that?

What does that mean?

Josh Davis: that mean?

So there are, we've divided for this
trial, and in fact, in, in reality,

it's divided this way as well.

Staph aureus into three different, uh.

Biological types based on its
antibiotic susceptibility, um,

to get into the SNAP trial.

The patient can have any one of these.

It's just that they're then categorized
according to what antibiotics

their organism is sensitive to.

Or susceptible to the most common kind,
which is around 70% of all is called

methicillin susceptible staph aureus.

So methicillin is an old fashioned
antibiotic no one uses anymore.

Uh, but it's still called
that for historical reasons.

Um, and so MSSA is, uh, the common kind
of stuff, staph aureus that is susceptible

to penicillin, family antibiotics.

Um.

Then there's MRSA methicillin
resistant staph aureus that Steve

mentioned earlier, which is more,
tends to be resistant to multiple

antibiotics, not just to methicillin.

Um, and has caused a lot
of problems with, um.

Outbreaks E either in
hospital or in the community.

So in the USA, for example, it
famously, there was a strain called

USA 300 of MRSA that caused outbreaks
in the community and including

among sporting teams, for example.

Um, and that causes a lot of
skin abscesses and sometimes

bloodstream infection.

So there's MRSA, which is
harder to treat generally.

There's MSSA, which is the garden variety
one, and then the third kind is PSSA

penicillin Susceptible staph aureus,
and that's like the OG old fashioned

one staph aureus that existed before we
started using antibiotics in the world.

So pretty much all staph laureus
was penicillin susceptible at one

time, and now, uh, between 10 and
20% of it is we can still use the

old fashioned penicillin to treat.

Scott Berry: Okay, so you refer
to these as, at least maybe in the

statistical plan, the as silos.

These are baskets of patients and every
patient fits in exactly one basket.

Uh, okay.

So these are mutually exclusive
categories of, uh, of it, um, uh, in that.

Okay.

So within a patient, in each of
those, you started the trial with.

Three domains, you call it.

And the three domains each have individual
questions and at the same time, you're

going to potentially have patients be
randomized to each one of these domains.

So they could get one from domain
A, one from domain B, and one from

domain C, one patient in that.

Uh, so tell me about the orig,
the, the first three domains.

I think the first one was the
antibiotic backbone domain, uh,

which varies depending on the silo.

You described the three groups.

So what are the questions
you're asking in that?

Steven Tong: So the questions asked in
this backbone domain is really, which is

the key antibiotic we should be using to
treat this particular patient with this

type of staph or, so if it's penicillin
susceptible, we're randomizing patients

to either stock standard, benzoyl
penicillin versus an anti staphylococcal

penicillin, which we call flu oxacillin
or oxacillin in some countries.

The second silo for MSSA in this backbone
domain, patients are randomized to

flufloxacillin, the anta copen, or to
a, what we call a sporin called zolin,

or cefazolin as they say in the us.

Josh Davis: No, they say

Steven Tong: then zolin, sorry, I,

Scott Berry: Okay.

Josh Davis: Zolin.

Steven Tong: uh, and then the, the
third silo for MRSA, we're actually

looking at standard therapy, which can
be an antibody typically vancomycin.

In some cases daptomycin,
but typically vancomycin.

And then the comparator is
vancomycin plus cefazolin.

So we have this backbone domain.

It depends on which silo the patient is in
as to which agents they get randomized to.

So we are actually answering
three questions in that domain.

And then the second domain
is to look in adjunctive

Scott Berry: Can I hang, can
I, can I go back to this?

When you're doing this, these are all sort
of separate questions, separate patients.

There's no statistical
borrowing across those three.

For the backbone to give within each silo.

So these are somewhat separate
independent questions statistically.

Right.

Okay.

Okay.

So then, sorry, I inter
I, I interrupted domain B.

Steven Tong: So domain B is
to look at adjunctive therapy.

So this is an additional antibiotic.

To what's already been given
in that backbone domain.

And in this case, our first
adjunctive antibiotic is an antibi

called Clindamycin, which acts to
switch off toxins that have been

produced by the staph orus bacteria.

Uh, so it has an antitoxin effect.

So we're randomizing patients to
either receive clindamycin or no

clindamycin for a five day period.

Scott Berry: Hmm.

Steven Tong: So that is a crosswalk.

Three silos.

Yeah,

Scott Berry: so that
could be a, oh, sorry.

That could be a valuable thing, even if
it's penicillin susceptible, that it,

there's still toxins and this is almost
as, as you call it, adjunctive, uh, to it.

Okay.

Um, and, and there is statistically.

Your, your, the analysis of this,
is it assumed to have the same

efficacy in all three of your silos?

Steven Tong: It is.

We were actually discussing this
last night with our statistical team.

We are pooling the patients
from the three silos.

But we have pre-specified looking at Mr.

RSA specifically as well as to whether
there's an interaction between domain A

and domain B for that group of patients.

Scott Berry: Yeah.

Okay.

And then, and then the early
switch domain, domain C.

Steven Tong: So I'll continue.

So domain C is looking at.

Whether it's safe to switch to oral
antibiotics at an earlier stage than

is typically accepted at the moment.

So usually we tr for years and years and
years we've been treating stat bloodstream

infections with something like four to
six weeks of intravenous antibiotics.

So you can imagine patients have to
stay in hospital for a long time.

They need to have intravenous lines
or what we call, you know, a PICC

line in for a long period of time.

We're looking to see whether it's safe
and e equally efficacious to switch

patients to oral antibiotics, meaning
that they can potentially get home earlier

and have intravenous lines taken out.

It's a bit of a dog that you should
continue with the intravenous

antibiotics the whole time.

So this is, uh, kind of challenge to
that existing dogma, but potentially

has major consequences if we can
show that it's non-inferior to IV

for patients and for healthcare.

Scott Berry: See I, so how much,
sorry, you may have said this.

How much earlier is the switch to oral?

Steven Tong: So there's actually
two points at which patients can

switch to oral antibiotics if they
have low risk uncomplicated disease.

This can happen at day seven and continue
for another seven days, either IV or oral.

If they have a more severe, more
complicated high risk disease, this

then happens at day 14 at which they're,
you know, deemed to be eligible if

eligible randomized, to then continue
IV for another either two to four weeks

or have all of that as oral therapy.

Scott Berry: Okay.

Okay.

So when patient comes in and they can
be randomized among all three domains,

uh, factorial, randomization, uh, to
that, does your trial do response?

Adaptive randomization?

Josh Davis: We, we don't at the moment.

Uh, we, we've written the core protocol
to be future compatible with our

ar, but, um, currently all of the
domains and all of the comparisons

only have two interventions, um,
that none of them at the moment

have three or more interventions.

Um, and so partly for that
reason, we've not used RAO.

We've just used simple
one-to-one randomization.

Scott Berry: Okay.

And frequent interim analysis in the trial

Josh Davis: Well, I dunno what,
what you define as frequent,

but, um, after every 500 patients
are recruited to the platform and

reach their 90 day follow up, um,
and, and, and interim happens.

So that's now happened so far, I think.

Um.

Uh, I can't, we're up to 6,000
over around 6,000 patients.

So, um, I think the 5500th patient
has had its interim analysis.

That's 11.

Scott Berry: Yeah.

Yeah.

Which, which I think for most
draws that's, that's a big number.

Um, you know, more than one or two.

And this is, this is a perpetual
trial that will continue on.

And so every 500 patients
you do an analysis.

Do we know any answers?

Do we know any answers?

Uh, your primary endpoint, we
haven't talked about the outcome.

What's the primary endpoint of the trial?

Steven Tong: It's 90 day
mortality, so pretty cut and dried.

Uh, we debated.

You know, should you do composite?

Uh, should you bring it
at an earlier time point?

Should it be all cause or
cause specific mortality?

So they are the kinds of things we
discussed, but 90 days just simple,

uh, unbiased, easy to collect, uh,
and very hard to argue against.

Did we find a difference
in nitrogen mortality?

Scott Berry: right, right.

And you, you do, I I, I mean,
you, you do have sample sizes

to address those questions.

I assume a huge issue in other
scenarios is your sample sizes

are too small for mortality.

But the global effort here gives
you that incredible opportunity

to, to, to look at that.

You're also enrolling, um.

Youth in pediatric it says, so it's
adult and uh, peds in the trial.

Josh Davis: Yeah, this is

a, this is one of the kind of innovative
aspects of the trial is that very few

trials in the past, at, at least in
the infectious diseases field, have

enrolled both children and adults
into the same randomized trial.

Uh, so most of the questions that are.

Are being asked in the trial apply
to children who have this, this

illness as well as to adults.

Uh, so that being, um, whatever age
a person is, if they have a staph

or is bloodstream infection, they
can potentially enter the trial.

Um, and then there is statistical
ways of dealing with children

and adults, uh, separately, but
with borrowing between them.

Scott Berry: Uh, yep.

Uh, and, and the US FDA just put out
a guidance document for, for Bayesian

analyses and all that, it highlights
borrowing from adult to peds.

And you're enrolling, you're enrolling
far fewer peds than adults, I assume.

Josh Davis: Yes.

Yeah, it's a, something like less
than 10% of the patients are, are

pediatric, um, at the moment.

So.

Uh, actually 258 children
out of nearly 6,000,

um, patients in total.

Uh, and as you say, uh, the a the data
from the adults can help inform the

pediatric, uh, in, um, statistical

testing.

Scott Berry: Mm-hmm.

Steven Tong: Part of our difficulty is
that, um, the primary outcome occurs

very, very uncommonly in children.

So in children, you know, mortality
rates probably sit around 2%.

Whereas in adults, um, it's as described
before, it's about 20% usually,

and that really increases with age.

So once you're 60, 70, 80, much more
likely to die than earlier age groups.

So that creates some issues with
how we do the statistics for this.

Scott Berry: Yeah.

Yeah, we should very much point out
you have a fantastic statistics team.

Uh, and you, you mentioned the,
the, the meeting of this group I

know from Barry consultants, Anna
McLaughlin was involved in the original

designs and modeling and simulation.

She's now unblinded and involved.

Carrying out interim analysis, so
thought we would, would not have her on.

And Dr.

Julie Marsh, uh, heads up your statistics
working group and, uh, lots of really

cool things going on in the trial.

So they do an a, a really awesome job.

So we wanna make sure we, we shout
out to, to, to their efforts.

Steven Tong: Hmm.

Josh Davis: Yeah.

Can I, can I just add, I was just
reflecting on this last night that one

of my favorite, um, aspects of this
trial is learning about the statistical

approaches and having.

Discussions and often debates
about, um, little statistical

questions that arise all the time.

And so Steve and I like are learning a
lot from the statisticians we work with

and, and really enjoy that, um, the
kind of intellectual challenge of that.

Scott Berry: Yeah.

Yeah.

It's, it's, it's a beautiful
integration of, of, of all,

of that, which is very cool.

So you have results in the trial.

So you've enrolled, uh, almost
6,000 patients and you've

had results in the trial.

What have you learned in Snap?

Josh Davis: So we've had three, what
we're calling domain conclusions so far.

Um, the first is in the MSSA
backbone domain that we've

defined what that is already.

But just as a reminder, that was comparing
cefazolin or, I like how the English

people say it cefazolin So comparing
cefazolin with flucloxacillin um,

for, um, methicillin sensitive staph
aureus, so the common staph aureus.

Um, so for that particular organism,
flu, oxacillin, or oxacillin is

considered, we're considering the
standard of care and the control

arm in the trial because that's what
most guidelines recommend be used.

In North America.

In fact, Zolin is more commonly
used even before the trial

for various reasons.

But there are doubts about
whether zolin is in fact as

efficacious, as flu oxacillin.

Um, and there are theoretical reasons to
think it might not be the preferred agent.

Um, and so for those reasons, for
severe infections, flu oxacillin

up until now has been recommended.

So it was a non-inferiority design.

Um, and we found that cefazolin um,
hit a statistical trigger for the

stopping rule for non-inferiority.

So the cefazolin in terms
of 90 day mortality was

non-inferior to flucloxacillin.

But not only that, uh, there was a,
a higher rate of acute kidney injury,

so toxicity from the antibiotics
in the flucloxacillin group.

Than in the cefazolin group
by a factor of about 1.5

times.

So, um, these.

Data have been, uh,
reported at a conference

at, at ESCMID Global, it's called.

That's the sort of largest infectious
diseases conference these days

that runs every year in Europe.

Um, and so the, the kind of ID
community are aware of them,

they've not yet been published.

They kind of at an advanced
review stage at a major journal.

Um.

So for, for that first domain conclusion,
it already has started to change practice.

Uh, definitely within SNAP sites because
we have a registry where we measure.

What's happening outside the

trial in patients with staph
bacteremia at participating sites, and

anecdotally elsewhere as well, but it
probably won't be able to really, uh,

officially change guidelines until it's
published in a peer reviewed journal.

Um, but the short version
in my mind is flucloxacillin

is dead for this indication.

Um, and cefazolin is
now the preferred agent.

Um, Steve, I'll let you talk
about the other conclusions.

Scott Berry: All right, so let's
come back to this a little bit just

to, so now you've got this piece of
M-M-M-S-S-A where you've got this

conclusion and now what happens to
a patient that comes into your site?

That's MSSA, the, the backbone is.

Now investigator decision on that?

Or does the trial
deterministically give them zolin?

What happens now in the trial?

Steven Tong: the patients, it,
it is investigator decision now

or clinician decision based.

We hope that most will be using Zolin, uh,
but we are not deterministic about that.

Scott Berry: Mm-hmm.

Meanwhile, that patient, so they
would not be randomizing and they

would not be contributing to that
question anymore, uh, regardless of

which arm they get because they're not
randomized, but they still could, would

contribute to domain B and domain C.

Steven Tong: Correct.

Scott Berry: Okay.

Okay.

So fantastic.

Um, that's the first result.

And, and, and generally, it sounds
like the community has been very

receptive, even though it's not out
in publication yet, that, that this

is, it's changed care, as you said.

Steven Tong: Yeah, we think it has,
and that's based on some of that

internal registry observational
data that Josh mentioned.

So patients who either are
not eligible, don't consent.

To the randomized platform,
we can consent them to just

observational data collection.

Uh, and so we can see in real time where
there's actually changed in practice

outside of the randomized interventions.

And in Australia at least, we have
noted that since our results have been

publicized, uh, there's been a real drop
in the use of flu oxacillin from something

like, you know, 50% down to about 10%.

Scott Berry: Is there a
cost difference of the two

Steven Tong: It's very minimal in some
areas of the world, so particularly

in Europe, actually, Razzo hasn't been
registered or isn't on the formulary.

So in fact, in the UK they had to
bring in Razzo onto the formula

for the purpose of the SNAP trial.

Scott Berry: Hmm.

Steven Tong: Uh, and so it will take
a bit more time, I think for that, for

these results to really impact because
there's, you know, greater forces at work.

It's not just the individual clinician
decision for whether to use zolin.

Scott Berry: Hmm.

Hmm.

Uh, that's fast.

So maybe even potentially your data could
be used to register zolin in the eu.

Josh Davis: Yeah, we hope
that's going to happen.

I mean, anecdotally we've heard in the
UK that, uh, steps are being taken,

um, to, to get the supply chain
working, make the drug available, but

we don't know, uh, officially what's

happened yet.

Scott Berry: Very cool.

Okay.

Uh, so result number two.

Steven Tong: So this is the PSA
silo for the backbone domain.

So our penicillin susceptible
staph aureus, uh, and here we were

comparing benzo penicillin, or
penicillin versus flu oxacillin.

And flu ox is once again
considered the standard of care,

and part of that relates to.

The laboratory determination of penicillin
susceptibility is not straightforward, and

many clinicians and labs actually don't
report out on penicillin susceptibility.

So this is a little bit of a leap
because we are saying we, we do

actually trust what the laboratory say.

Um, and we can use that to influence
what treatments should be provided.

So patients who have been carefully
validated to have penicillin

susceptible staph orus were randomized
to penicillin or flufloxacillin.

Uh, and in this case we had a small
patient numbers, so it was about

280 patients, whereas with Ms.

SSA it was 1400 patients.

So as a result, we have less power
to, to demonstrate, you know,

very clear conclusive findings.

But we saw very similar signals, so
much more toxicity with flu oxacillin.

Uh, the rates of acute kidney injury,
uh, were almost twice as much with flu

oxacillin compared to benzo penicillin.

Um, and I think that met,
uh, uh, a superiority.

Um, I iis, uh, in terms of the
posterior probability of more than 0.99,

that it was in fact superior
penicillin, superior to flu

clots, and on a mortality count.

Uh, so the primary outcome was also better
with penicillin compared to flu oxacillin.

Uh, we didn't quite reach a trigger
with that because of the smaller

sample size, and it's worked.

I guess taking a step back as well,
the reason this, uh, domain closed

was that the toxicity signal for a KI
for acute kidney injury with flucox

IC was seen in both the MSSA and the
PSSA silos, and therefore the data

and Safety monitoring committee.

Asked us to stop the trial
because of that, principally

because of that toxicity signal.

Now the non-inferiority trigger was
reached in the um, razzo versus flu

oxacillin comparison, partly because
the larger sample size, but with the

only two 80 patients with PSSA that was
not quite achieved, but the very, very

similar directional toxicity signals.

Scott Berry: Hmm.

So, so that's interesting.

So a, uh, uh, it strikes me
that the MSSA and PSSA are,

are almost similar type groups.

They're not, they don't have
a resistant bacteria at that.

So if their penicillin.

Um, susceptible.

You give them penicillin.

If they're not, they're MSSA, does
that mean they're in that bucket?

This, this sort of, and then they
might get ke zolin in that case.

So it's value.

Does that mean it's valuable to
test for penicillin susceptibility?

Steven Tong: So I think it, it is, and
what the trial definitively shows, I

think is you can trust that test result.

So many clinicians would, even
if they see it from the lab,

would not trust that result.

But what we've shown is not
only can you trust that result.

Then in fact, penicillin is almost
certainly better than flu oxacillin.

So these are patients who
previously would've just gotten

flu oxacillin because the testing
was either not done or not trusted.

Scott Berry: Hmm.

Hmm.

Fascinating.

Okay.

Uh, so that, that might even change
diagnostic criteria that are used before

decisions are made, uh, at different ways.

Would you ever compare penicillin to ke
zolin in the PSSA group or the m uh, MSSA?

Presumably if you're in MSSA, you don't
test as being sensitive to penicillin.

Josh Davis: Yeah.

Yeah.

So

MSSA partly.

You know, by definition
is penicillin resistant

Scott Berry: Hmm.

Josh Davis: that, not
ms, but in PSSA, yes.

You put the, hit the
nail on the head there.

The zolin versus penicillin
is an unanswered question.

Um, the one that's, that's.

Kind of been brought up
after this conclusion.

And in theory, yes, we could test that.

We've been sort of discussing
and debating should we do that?

Is that a good use of resources
in the ongoing SNAP platform?

Um, and that's not quite resolved yet.

So

we, we haven't actually directly compared
those two antibiotics, but both of

them we know will work well for PSSA.

Scott Berry: Hmm.

And, and just going back to your
results, these are, for, both

of them are specific to adults.

Steven Tong: Yes, that's right.

Yeah.

So,

Scott Berry: you, you haven't had a
conclusion in peds yet, is that right?

Steven Tong: Correct.

And um, and so we actually have
reopened that domain for peds.

And in for peds, we are actually
combining the PSSA and MSSA, we're

putting them together and the comparison
is now flu oxacillin versus razzo.

And for that entire patient,
those two patient groups now.

Scott Berry: Hmm.

Alright.

I, and I, I didn't miss one.

Those are, those are two
con your two conclusions was

there, has it not been a third?

Josh Davis: Uh, there,
there has been there.

Yeah.

There has been a third, but we
don't, uh, we haven't actually

reported the detail on that one

yet.

But What, but what is in, I guess the
public domain is that the platform

conclusion has happened, um, and that
the DSMC, the Data Safety Monitoring

Committee has told us to stop enrollment,
um, due to having hit a futility trigger.

So that was the, the adjunctive
clindamycin domain that we mentioned

previously, where, where people were
randomized to get clindamycin or not for

five days in addition to whatever their

backbone antibiotic was.

Um, and that after around 4,000
patients, uh, were recruited, were

randomized, um, has hit a futility
trigger, um, that we're never going to

see a difference in 90 day mortality.

W by giving Clindamycin.

Um, so that's about the amount of
detail we know thus far, and we

will be presenting those results
in detail at the Es mid conference

again this year in, um, in April.

Scott Berry: Hmm.

Fantastic.

Uh, what, uh, so the, the
engine that's generating these

results is, is just awesome.

What's the, uh.

What's the future now?

So you've got this platform and you're
interested maybe ke zolin to, uh, and,

and I know you say it four different
ways, so, uh, uh, and penicillin.

Maybe there's a comparison in, uh,
in those groups, but what other

things are you doing in the platform?

What's the future?

Steven Tong: So our priority is to try
and complete still ongoing, uh, domain.

So for MRSA backbone, we are up to
about six or 700 patients there.

We'd like to try and get
to an answer for that.

The early oral switch continues.

That's up to about 1200 1300 patients.

You know, as far as we know,
that hasn't hit triggers yet.

Okay.

then because we do have the platform,
we'd like to introduce new domains and

there are a number which have been funded.

Now some of these are heading
more towards, I guess, more

experimental therapies perhaps.

So one of them is serial phage therapy.

Uh, so phage, uh, viruses than infect
bacteria and can, lies the bacteria.

And when the bacteria get lies,
the bacteria fas or the virus

gets released and goes to infect
the next staphylococcus aureus.

And they're very specific and directed
to the bacteria and not any host cells.

So phage therapy used to be used in
particularly Russia many, many years ago.

But with the advent of antibiotics, um,
phage therapy really fell into disuse.

Scott Berry: Hmm.

Steven Tong: But with increasing
antimicrobial resistance across

the world, there's been renewed
interest in use of phage therapy.

So we'd like, we, we have funding in
Australia to introduce a phage cocktail to

patients with staph bloodstream infection.

So that's one new funded domain that will
bring on board during the course of 2026.

There are also domains to look
at, um, adjunctive, antiplatelet,

or anticoagulation agents.

Uh, so there's some interesting
preclinical science suggesting they

may be of benefit, but obviously
there are also risks involved in that

they could increase risk of bleeding.

Uh, but we are going to do small,
almost really phase two trials.

So instead of thousands of patients
somewhere in the border of 300

patients into those domains.

There's a diagnostic
strategies domain as well.

So we're looking at, uh, PET CT scans.

So that's, uh, nuclear medicine imaging
to look for hotspots of infection.

There's observational data to suggest
that doing these PET scans gives you, uh,

more information that allows you to tailor
the duration of your therapy better.

So we'd actually like to
randomize patients to either

PET CT or no PET CT scan.

And determine in a randomized
fashion whether this observational

data holds up or not.

Scott Berry: Hmm.

Uh, diagnostic, diagnostic things
like that are strikingly hard.

So it's, uh, it's, I, I compliment
you for taking that because

a lot of times the patient.

Is treated exactly the same
regardless of what the scan, what,

what the diagnostic thing says.

This has been, you know, a plague
for cancer for a long time is

well, that you do the same thing.

So it's really hard to show clinical
benefits so that, that's awesome.

It's kind of fantastic because you've
shown presumably that penicillin,

penicillin susceptibility is
already a good diagnostic thing

to add, so that, that's awesome.

Steven Tong: Hmm.

Josh Davis: I might just add as a
general kind of comment about these new

domains Steve's talking about, and he
is only got through about half of them.

So far, um, is that nearly all of them
are not arising from myself or Steve.

They're arising from our, the, the network
that we've built up over, over time.

Um, and that's one of the great things
about Snap is that there is now a network

of hundreds of investigators in, in the.

13, 15 countries.

Um, and a lot of them have brought
their own ideas to the platform.

So for example, the, um, the
anti-platelet domain that's being

led by an investigator from Spain and
another one investigators from Canada

are both doing anti-platelet domains.

The bacteria phage is, um, other
colleagues from Australia and Netherlands.

There's a, there's a phosphomycin
domain, which is another adjunctive

antibiotic from Germany, et cetera.

So, um, a lot of snap is we kind of
like planted a tree and now the tree

is like growing and running away from
us and getting a life of its own.

That's, that.

was a bit, that was a
bad metaphor, wasn't it?

But you know what I mean?

Steven Tong: But, but what's
nice is that the funding isn't

all coming from Australia now.

Like each of these investigators has
applied to their local funding bodies to

support, um, both entry of their patients
into the overall SNAP infrastructure,

but also to support their specific
questions that they would like to answer.

Now, whether.

Recruitment to those new domains will be
country specific or where they'll actually

use the whole of the SNAP infrastructure.

I, I guess, is, is we'll wait to see,
but the attraction is that once you

write a local protocol, we can try
and operationalize that such that if

there's other countries in the world
who are interested in that question,

they can also recruit to it and help
you get to your answer more quickly.

Scott Berry: So we've always described
that once you get the adaptive platform

trial up and running, like you have to
add these additional questions to bring in

these investigators with these questions.

The marginal cost, the marginal lift
is so much smaller than had they.

To design their own trial
and go get their own sites.

So you're asking more questions
than you ever would've, and

the marginal cost is less.

Is that a fair statement that you've
learned or, or am I overhyping this?

Josh Davis: No, I think

that's

Steven Tong: think it's the.

Josh Davis: true.

That is a fair statement.

But it do, there are, there
are, you know, some downsides.

There are a lot of added complexity
that are hidden and uh, I guess

as an example, we we're trying to.

Keep with the principle
of a single database

for electronic, um, data capture, um, and
randomization and eligibility screening.

That all happens in a single database,
which is fantastic, but we now have a

long queue of, um, of domains and things
to be added to that database, and, and

that takes time rather than, you know,
building individual databases would

be faster, but a lot more expensive.

Scott Berry: Mm.

Yeah.

And, and you've got, all of these
have, uh, there's central cost to

the trial and then there's cost at
individual sites, there's governance.

Uh, so, so I'm sure you're, you
have lots of things you're trying

to tackle that goes with that.

But that, that, that is very cool.

So, um, the, it seems like the adaptive
platform trial itself has accomplished

a lot of, uh, additional goals.

Any lesson learned that you'd
like to pass off to people?

Uh, I, I don't know, regrets is
maybe the wrong question, but general

lessons learned from jumping into this.

Steven Tong: Look, I think we've
learned heaps just by doing this.

So Josh and I had run.

Uh, maybe, you know, three
or four clinical trials.

Much, much smaller in scope than this,
and this is really just snowboard.

So we've learn a lot in that process.

I think there are some things we
would do differently, particularly

relating to the database, um, design.

I think we've made it a bit
too complex with lots of

internal logic in the database.

That's meant that any adaptations
or changes, new domains that we are

trying to introduce now, there's a
lot of downstream, uh, unintended

complications because of how
we've initially designed things.

And I think if I were to go back, I would
try and simplify that initial design.

It's great for sites, uh, because the
logic means you get to skip questions

that are not relevant to your site or
to the domain that you're recruiting to.

But some of that, as I say, has created
complications down the track now, which

costs money and time to try and unpick.

Uh, and we could have avoided that.

Yeah.

Scott Berry: Mm.

Josh Davis: Uh, one, one other
difficulty, which this is not

something I'd do differently because
we don't have control, is that.

Funding cycles, funding agencies,
generally government funding agencies,

um, fund trials for a number of years,
three years, five years, for example.

And as you mentioned, Scott,
this is, we've set this up as

a perpetual trial potentially.

So as long as funding exists
and there are still questions to

answer, the trial can keep running.

Scott Berry: Hmm.

Josh Davis: However, it's really
hard when you're pitching to

a government funding agency.

Um, you can't say it's perpetual.

You have to say you're going to
achieve something in five years.

So we've been lucky enough to be able
to get a second, five years of core

funding from the Australian government.

Um, and then also in, in many of the
other participating countries, they've

had funding from their governments,
but it's a lot of work to keep going

back and, and getting a funding again.

So it would be really interesting to see.

I, I think just funding
agencies have not caught up.

With,

um, this way of doing trials, right,
which is still relatively new.

Scott Berry: Yeah.

Yeah.

I, that's a whole new podcast of, uh,
that whi which in some ways you, you're

starting to almost do this learning
healthcare, where at the 150 sites you

have globally, you're learning things.

It's implemented probably
better and faster.

Now you add new questions and you're
doing what we originally said.

The goal of this was.

Was to improve the outcome of
patients with staph infection,

which is was fabulous.

So going all the way back to
Darwin, I feel like you've improved

the outcome of patients in Darwin
through these efforts, I imagine.

Fantastic.

Steven Tong: so that's, that's
what we're looking for and I keep

reminding our teams and all the
investigators at sites that ultimately

this is to improve patient outcomes.

Like we're solving the weeds of
the everyday operational aspects.

It's easy to lose sight with that bigger
picture, that because of what we've, what

we've done and our community has done.

I think less patients will
die from this disease and less

patients will get toxicity from
flucloxacillin at this stage.

And we'll, we'll find
out new things as we go.

Scott Berry: Yeah.

Fantastic.

Well, I, I want to compliment you
both on leading this effort, and I

know you, you'd be the first one to
say there's huge teams and efforts of

lots of people, a, a global effort.

But this is such a, a, a really neat
example of, of, uh, adaptive platform

trial and a global, uh, action on this.

So congratulations to both of you,
and thank you for joining us here.

Josh Davis: Thank you, Scott.

Thanks for having us on.

And I guess thanks for, um, for Barry
consultants as well who have been

absolutely instrumental in, in getting
this set up and getting it to work.

Scott Berry: Yeah.

Yeah.

And again, compliments to
Anna for all her work in that.

So, and uh, everybody for listening,
appreciate you joining us.

Until next time we'll be
here, uh, in the interim.