Selen Dar

Muscle-Building Workout and Diet

>>Paul Thomas:
Our next speaker is Dr. Mary Hardy.
She’s a physician; she is going to be speaking
on the concept of safety, and its measurement
and its reporting. Dr. Hardy is board certified
in internal medicine, a specialist in botanical
and integrative medicine. She has, really, a lot
of extensive experience in evaluating the evidence base
for the efficacy and safety of complementary and integrative
medicine practices. She’s currently affiliated with
the George Washington University Medical School’s
Integrative Medicine program. This is Dr. Hardy’s first
presentation at our practicum, so please give her a hearty
and warm welcome. [applause]>>Mary Hardy: Hi everybody. I first would like to thank
the organizers for inviting me. It’s nice to be here, it’s nice
to see what a practicum is. It’s certainly been eye-opening and very practical
for me as well. So, I’m going to talk
to you about safety. And luckily this program
was very well designed, because I think
that there’s been a number of
complementary presentations that touch on each other,
so that I was afraid I would have to give you
the entire context, which would have been
half of yesterday and everything that happened
this morning. So, happily, I don’t have to
do that in thirty minutes. Ultimately,
every assessment of you, whether or not to use
a supplement, drug, whatever, comes down to a
balance of risk versus benefit. Now, I’m not going to
talk about benefit today. Johanna’s talked a lot
about patient values. Benefit’s already been
touched on as well. But this issue of risk is,
I think, a critical component, especially when the benefit
is either not very pronounced, not thought to be
very pronounced. But it’s important to also know
that risk varies depending on which chair
you’re sitting in. So, risk for a regulator
will be different from risk for a patient
who may feel more compelled about the disease entity
they’re trying to treat, than perhaps a risk
as assessed by a health agency. And, ultimately, most safe use
of these things comes down
to an appropriate interaction between issues related
to the patient and issues related to product quality.
Yesterday, Adam talked a lot about all the issues
around standardization, evaluation of product,
and product quality. And the issue of looking
at dietary supplements, but especially
for herbal products, is a critical issue. And so, before you do
any other assessment, you want to make sure that
what you think you’re looking at is what you’re actually
looking at. Now, patients and–I call
them patients, most other people
call them consumers–and therein
is part of the problem; many products would be used
without supervision, without appropriateness for use. So, for example,
someone who has hypertension, who is taking a product that may affect blood
pressure adversely, again, not always the best mix
of patient and product. The product itself
may be made well and without inherent risk,
let’s say, but in that patient
it’s an inappropriate choice. So, we talked about one of the
medical issues they may have, but also stage of life. Metabolism varies
depending on your age and your other medications,
et cetera. In looking at the adverse
event reports for ephedra, which we’re going to talk about
at the end of this presentation, it made a huge difference
what state the person was in to begin the adverse
event report. So, being dehydrated,
in actively in exertion, et cetera.
External environment. It was a very different matter
taking this product on a cool fall day versus
in a hot sauna, for example. And we’re going to talk
a lot subsequently– Dr. Gurley’s going to talk about
interactions and supplements that allows me
to skip that point. There also can be some racial
or ethnic variation, both on some of the SNPs and the way things
are metabolized in patients, but also into other activities
patient populations may or may not use,
as Johanna had already told us. Now, if you’re going to look
broadly at quality issues, this is the list. And, again, we’re going to talk
a little bit about some of these. Normal but toxic
doesn’t happen that often. Now, when we say
intentionally added, that’s usually drugs,
not toxins. Contaminants–
and there’ll be times when the manufacturer
made a mistake, when they just
put in ten or twenty or sometimes a higher amount
of any substance. And the reason why we want
to kind of look at the sort of
the whole picture of this is because the remedies
are going to vary. If it’s a mistake
in manufacturing, fine. If it’s a contamination,
that’s one thing. Inappropriate patient
population, inappropriate labeling, that’s what the FTC is for,
et cetera. So, you kind of want to start
to take globally the things I’m telling you, and things–
start to triage them based on, “what would I do
with this information? How would I make
an intervention that would provide
a greater safety environment for my patient?”
Our beloved Norman Farnsworth, who is not with us anymore,
has said, and I believe
this is actually true, that if what’s on the label
of an herbal medicine and dietary supplement
is what’s in the container, toxicity rarely is encountered. And so, again,
that critical first step, getting what you think is in
the container in the container. Adverse events can come from
a variety of different sources. We’ve talked a little bit about
adulteration, inherent toxicity. These substances can have
a pharmacological effect which is minor,
it may be a direct action, it may be mediated
through the CYP– the liver enzyme
detoxification system– it can interact with other
drugs or dietary supplements. Patients can misuse
any product, they can double,
triple the dose. And then, after all
that’s said and done, people just do weird things, which is medically known
as idiosyncratic reactions. So, contamination.
You’re going to notice, if you’re following
along with my slides, that was my first cut
at a really low dose of slides, a small number of slides,
but I’ve gone further than that. So, you’re going to have
more slides in your handout than I’m going
to talk about now. Those are there mostly
for another example, the same thing
I’m talking about, additional papers to bring to
your attention, or definitions. There’s a whole bunch
of definitions that I didn’t figure
you needed me to read to you. All right, so contaminations. This is looking at a study
that reviewed the literature and looked at a number
of reported adulterations. Now, there are
unintended adulterations, and some of these things,
seaweed for example, can have heavy metals, they’re just present
in the water. And calcium supplements, especially those that are made
from oyster shells, can have heavy metals
in them. Again,
environmental contamination working into the basic
product source. This, perhaps, is a bit
more problematic just because it’s evil-minded.
You have economic adulteration where you take something
that’s well known, and then you substitute
a cheaper or different compound
for that. And some of it
is a different herb, some of it’s putting
in a compound that will make
the marker compound look higher
than it really is. We had a whole spate
of reports in, especially on the West Coast, Chinese patent medicines
were contaminated with drugs. If it was a patent medicine
used for pain, there might be
an actual pain medication, either over the counter or a usually prescribed
medication adulterating. The highest risk products
for intentional adulteration are erectile
dysfunction products, which have a lot of
sildenafil analogs in them, performance enhancement, which will often
contain steroids or other banned substance,
and weight loss products. These are the highest
risk products for contamination. Misidentification
can often be an issue, and there’s a couple
of examples in your handout. This is one that I thought
was particularly interesting. I was serving on
the Natural Health Products of Canada advisory panel when these six adverse events
were reported. And the basis of these adverse
events, a question came up, whether or not to label
all black cohosh products as potentially
being hepatotoxic. Well, going back
to my first principles, I asked that we would assure that what we thought
was in the bottle was actually on the bottle.
And it turned out, it wasn’t. It was, instead
of an Actaea species, it was Chinese black cohosh,
Actaea foetida, instead. So, these did not have
the product in it, so therefore labelling
all black cohosh products, which were to supposed to be the
American version of black cohosh (Cimicifuga Actaea racemosa),
as having a problem, when in fact
there were constituents that were more commonly found–
natural but toxic constituents that were normally found
in this herb that were not present
in the material that actually got
in the bottle. So, it was cheaper to buy
a “Cimicifuga,” a “black cohosh”
from China than to use
the actual correct species. So, that’s just a quick dash
through some of the things that you’ve heard
a little bit about, looking at it specifically
from a safety issue. Now, if we’re going
to make an evaluation, we have to get our information
from someplace, so where can we find
useful information? So, the first place
I like to start is to talk to other people who have had experience
using a particular herb. They often will know
astonishing amounts about the herbs
you’re looking at. And that includes working
with native traditional healers in several different places
across the globe, and most closely in Peru
and the Amazon. And this provider, this healer,
this herbal expert, would know things
about the plants that were
just astonishing to me. Then you want to have
a sense of, “have there been pharmacological
or mechanism of action studies?” That may give you
a reason to suspect that something might be
the cause of an adverse event, or might potentially lead
to an adverse event. What kind of toxic
constituents might be present? Look at pharmacovigilance
reports, people publish things
in conference abstracts, don’t complete
observational data. But the biggest chunk
of what we look at is adverse event reports, and that gets a lot of weight
put on it, and it gets a lot
of interesting attention. There are adverse events
reported in clinical trials: controlled, randomized,
case-controlled trials. But also just, out of the blue,
something really bad happened, and this is what
the person was taking. Sort of post-hoc reports. There’s a very
small pool of data, but mostly there’s not
enough data to begin with, so there’s rarely enough data
to actually do a pooled analysis for risk
or adverse effects, so– And we’re going to go through
these topics in more detail, because, depending on where you’re getting
your information from, each of them will have
particular risks and benefits, limitations
associated with them. The first thing to recommend
to your attention is that people who are
a committed herbal user, committed dietary
supplement user, are going to be
very unlikely to report. Reports in general
are in the single digits, and if it’s a committed user
who doesn’t want to make trouble for the product
that they like, then they’re not
going to report it. And they’re less likely
to report an herb than a drug if they’re a committed user. So, most reports
that were made are not severe. Most of the ones that get
the most attention are, in fact, very severe
but they’re quite rare. So, if you just go to
the medical literature, because that’s your
sort of second impulse, “just go look it up, someone
must have written about this,” it’s actually quite difficult
to find this information. It’s not housed very neatly
under a single meSH term. If you use adverse,
adverse effects, complications,
poisonings– to get to a fairly
high degree of certainty that you identified
most of the events that you’re going to want
to look at, you have to put together
multiple terms. And, even at that, you’re going to have to whittle
away a lot of other stuff. So, just finding these events in
the literature is not that easy. So, let’s go back
to our historical reporter. Why can’t you just talk
to someone who’s had lots of experience
using an herb? Well, you know,
if someone’s taking an herb, whatever the law says
that you can use an herb or a supplement for,
people mainly still don’t hear maintain a healthy
mental status, they hear depression when they think about
St. John’s Wart, for example. So, the question
that was asked earlier about, you know, efficacy and “you can’t use it
for label claim but you have to test
for something,” that’s a very apt comment, and patients tend to make
that translation. So, many of the things that we
are being treated for here in a Western, First World society are nothing
like the traditional societies. They didn’t really have
atherosclerotic vascular disease in most of these countries
and most of these societies because people died earlier of
something completely different, or they had a completely
different diet, or completely different
exercise regimen. So, but there are times when,
between traditional and Western, there can be tight concordance. I did an analysis
on looking at Ayurvedic herbs of use for diabetes,
and in fact the concordance between the diagnosis
in Ayurveda, Madhumeh, and what we call diabetes,
was quite close. But, you have an easily
measurable outcome that you could do even in
a low technological society, so diabetes mellitus–
mellitus means honey– the sugar gets sweet
in a diabetic. So there’s a way
to make that diagnosis and follow that patient
in a low technology system. So in general,
this kind of safety data is best for acute toxicity. If you give someone a poison,
you know pretty quickly, right? But for chronic toxicity,
or unusual toxicity, not very, very useful. Then we take
a traditional preparation and we substitute an herb,
or we refine how it’s taken. So, chaparral is a bush you’d see in the west
of the United States. It’s got a very oily,
resinous texture because it’s in the desert. So, if you just drink that
in a tea, you’re not going to extract
a lot of the key ingredients because they don’t dissolve
very well. But take that same thing,
and either take a whole herb and put it in a capsule,
or extract it with alcohol, and you’ve now
significantly changed the traditional method
of preparation. And, in fact, you’ve actually
significantly changed the toxicity of that. So, you have to be aware
of that, when you’re– because everyone wants
to borrow traditional data for the safety
of their intervention. And that’s fine, as long as
it’s an equivalent analysis. When you look at
clinical trials, most clinical trials don’t
collect adverse event data, and they don’t
collect it very well. They’ll do some testing to look
for unexpected adverse events, routine labs, et cetera. But, there’s a section
always someplace, two or three sentences
in the paragraph says, “the product
was well tolerated.” So, let’s assume that your trial
is successfully randomized, not a benign assumption, and that you’ve got
two balanced groups. Then, if something
does show up, more often in the verum group
than in the placebo group, you can reasonably say, “yep, that can be attributed,
causality can be attributed.” But it’s, again, not usually
thought of very carefully when clinical trials are run.
And part of my pitch today is, those of you who are doing
clinical trials, pay a little extra time here and you can derive great
benefits for the rest of us. The most expensive part of
a trial is finding the subjects, randomizing them,
and following them. So, if we can just have you
start to think about adding in a more
robust safety assessment, doesn’t cost that much more, you can start to really build
this progressive database, as was talked about
in the last presentation. Now, so clinical studies,
randomized control trials, these are our gold standard. Why can’t we just do this
for everything? Why do we have to have
any other kind of study design? Why do we even care about
these adverse event reports? Well, it’s because
most clinical trials are very poorly powered
to find unlikely events. So, a clinical trial
will find a common event that occurs frequently, that occurs shortly within
the initiation of a medication, for most trials
are pretty short. But most trials
aren’t maybe long enough to find a long term effect. But the biggest problem is that
if you want a guarantee that you’re going to find
something that happens once in a thousand times,
which is not that rare, you need to enroll three times
as many subjects. So, the idea of finding
an uncommon, or even a not-so-common
adverse event in the clinical trials,
largely going to not happen. The numbers are against you. Here again, the specialty
population that was discussed, under-reporting by observers,
how you get your reports. And then there
can be event bias: how you’re questioned, how you’re asked about this,
et cetera. So, this is one of these
long slides that tells you all the ways in which we generally
don’t do things very well. But, specifically
for safety studies, you know, the drug was generally
well-tolerated. Well, what does that mean? Some people have a little bit
of stomach upset, only one person
really threw up a lot. You know, what? So not being specific,
not providing the actual data, not providing
the actual numbers, in general, just not being specific enough.
And there are two kinds of ways to elicit
these adverse event reports. The one is, researchers
will just wait for someone to spontaneously report
something to them. Spontaneous reporting is
generally very under-reported. I mean, because anything
can make you feel a little wonky for a couple of days.
It’s difficult to assess. On the other hand, if someone
specifically starts to ask you, “did you have this, did you have
that, did you have that?” that can predispose you, can bias you towards
having those side effects. I was included a in hepatitis
vaccine trial as an intern, and I came back
to the study examiner after two weeks and said,
“God, my arm’s been killing me! And I felt like I had
the flu the whole time!” And the guy looked at me
and goes, “yeah, you were in
the placebo group, so.” [laughter] So, okay, you can really suggest
things that’ll happen to people if you aren’t careful about
the way in which you solicit. But, I do think some combination
of passive and active should really be included. And this just sets you
towards reminding you that there are CONSORT
guides to reporting, that we want to be sure
that we can generalize, that what we do is valid,
et cetera. So, again, just pay
a little bit more attention to this aspect
of your trial work. Now, looking at adverse event
case reports, which is where the biggest bulk
of this analysis comes from, and where most people get bent
out of shape about stuff. First of all, if you look
at the adverse event reports, you’re immediately crippled
because most of– Remember, I told you, I think the first move
you have to make is, “is what’s in the bottle
what you thought it was?” And most of the data sets
for looking at a large number of adverse dietary
supplement events. I looked at 80 event reports, and they actually did not have
an adequate description. One. One had an adequate
enough description that I could even be sure
of what they thought what was in the bottle
was in the bottle. Rarely, if ever,
is the material ever recovered, collected, and tested.
So, this is in an individual– it becomes a very big problem,
and as you go forward. Pharmacovigilance is just a name for a sort of
post-marketing surveillance. It’s usually a voluntary
process, very under-reported. It cannot prove causation
under any circumstances, even if you have
ten thousand cases. It’s technically not
a proof of causation. But it will raise
your index of suspicion and identify something that you might want
to pay more attention to. Causality can be proven,
maybe in a case-control trial, maybe in a randomized trial,
with the kind of limitations I already told you about. Causality,
at the end of the day, becomes the name of the game. Someone reports something
terrible that happens, what changes that
from a “sorry for you” to “uh-oh, we better really
think about this some more?” kind of an event. And that is what they told
you was the adverse cause. Is it the adverse cause? And how do you go about
assigning that causality? In your set of definitions,
there are two scales that will tell you
sort of the formal way that this causality is assigned, but I want to walk you back
through some of this stuff. And again, think about the kind
of data that you’re looking at so that you have a realistic
expectation for what you can and cannot get out of the kind
of data that you’re looking at. Randomized control trial, that’s
the point of the whole thing. Things that happen in one arm
don’t happen in the other arm; if it’s well randomized, should be able to be attributed
with all the caveats. Cohort, maybe.
Case series, nope. The only kind of adverse event that points towards
real causality for the material
you’re interested in is something called
a dechallenge/rechallenge event. So, let’s say I tell you
I got sick to my stomach from taking a dietary
supplement. That’s my report. So, you tell me, “take away
that dietary supplement, go about your business,
tell us what happens.” Well, my stomach upset resolved. Now, that’s kind of interesting,
still not causality. So, let’s say, because you’re
this kind of a person, you know, you really just like
to poke your finger in things to make sure that what you think
you think is what you think. You decide, “I’m going to start
taking that again, just to see if it really did
cause my nausea.” Sure enough, you take it again
and the same symptom comes back. All right, that’s the tightest
kind of causality you can get out of
an adverse event report. Now, immediately you can see
there’s certain problems. If death is your
adverse outcome, I can’t rechallenge you. [laughter] If severe heart attack or stroke
or seizure is your adverse event, I’m not going to want
to rechallenge you. So, automatically,
for the more serious outcomes in adverse event
reporting from case reports, you’re going to be
severely limited in being able
to assign causality. And there’s a bunch of other
different kinds of studies, which kind of amount
to sort of some post-marketing, some other kind of case
control studies, retrospective
historical controls, all that kind of stuff.
You can’t attribute causality. Now, if you are going to start
to try to attribute causality and you want to see
how tightly you can get this linkage
to link up together, you’re going to want to know
how severe this is. Obviously, if it’s minor
and self-limiting, goes away after a little
while whether you stop or don’t stop your supplement, or your dietary supplement,
or your herb– it’s important
that it’s reversible, that’s pretty helpful. Because, if you’ve had
an adverse event and it’s a completed
adverse event, then there’s less way
to intervene. You want to make sure
that the pharmacology or that the exposure linkage
with the adverse event is appropriate
and seems logical based on what you know
about your substance. In the ephedra
adverse event reports, and I read hundreds
and hundreds of them, people would say, “I took this, and it caused me
to have a stroke.” “Well, when did you take it?”
“It was a week before my event.” That’s, you know–
this as a direct causal agent, no additional doses
in between, that doesn’t match
with pharmacology. But if we have someone
that says they took something and the onset
of the side effect or the onset of the acute event
was appropriate for the pharmacology
and pharmacokinetics, that makes this a lot
more likely to be true. The one that really gets
everyone into a lot of trouble is this one: the absence
of alternative explanations. Now, this usually requires
someone with medical training to make this assessment,
because– and it’s classic, classic, classic in abnormal
liver function tests, liver toxicity
adverse event reports. Pretty much everything,
including the party last night, can cause your liver function
enzymes to go up mildly, okay? Couple of drinks
can really do that by the next day
of measurement. So, and Tylenol, there’s lots of
drugs that can affect the liver. So, trying to rule out
other events that could cause
liver abnormalities is quite challenging, and it takes very thorough
reporting to really show that there is no other
alternative explanation. The more explanations
you eliminate, the tighter your relationship
gets to be between the intervention that you
think caused the problem and the observed problem.
So, keep that in mind as well. Mostly, when we start
to look at this stuff, just even getting
the information, we don’t often
know everything. We don’t often have
all the information. We may not know what to ask
because we’re not clear what the mechanism of action
of a product is, et cetera. And there’s a lot
of variability, both from person to person,
and there’s– so, we can’t like
extrapolate very easily. I saw one person
that had this happen to them, therefore everybody
who’s like this person, the same thing’s going to happen
if they take the same product. There’s a lot of commercial
variability in the market, you’ve heard a little bit
about that already, conditions of use might vary,
composition of product, and the individual patients,
even if they look the same, often aren’t the same.
So this, again, makes it hard for you
to get that tight linkage. Pharmacovigilance in general,
this is what it basically is: it’s a voluntary program,
very rare events. This is the only way
can find them. And there could be static–
and this is for your slides, so you’ll know where you might
go to look to report stuff. It can be dynamic. So, for example,
people report in real time, either in the emergency room
as you come in. But there’s also new mandated
reporting–a new law about four or five years ago
put into place, reporting to the FDA for any reported side effects
of dietary supplements and over
the counter medications. So, that’s been quite helpful
and I’ll talk to you a little bit about
that kind of data in a second. So, post-market surveillance
can be disorganized, or non-existent,
or informal, or it can be quite
an organized process. And in Europe,
it is an organized process for dietary supplements there. You have to do a formal
post-market surveillance study. This is nice because it, again,
gets into the real world and the patients doing
what patients do, and it gets you to populations you might not normally see
in a clinical trial, and it gives you
kind of a snapshot of what the real world taking of
this medication might look like. Okay. This is
the new legislation, in case you really want to take
down that number and read it, but what it basically is — is it’s mandated reporting
through the same MedWatch services,
everything else. And having looked
at that data so far, we’re trying to do
a series of — look for trends over numbers, looked at the first
two years of data. We’re trying to get FOIAs
for the last three years. Anyway, so, just in general, there weren’t that
many events reported, even with mandated reporting. Given the millions
and millions of doses, the reporting levels
were relatively small. So, one thing we’ll see is, as this law
has gone into effect, have reporting numbers increased
as people got more used to this? And again, herbal products were not proportionately
reported more frequently. Herbal products we’re
a little bit picking on because they’re complicated
and they’re more likely to potentially be adulterated,
or some other kind of a problem. People are reporting, but voluntary reports
still represent a significant number
of adverse event reports, so mandated reporting
is not capturing everything that’s being reported.
We did see one signal, which was also acted on
by the FDA, and it was
a Hydroxycut product, so that one was taken out
and was found. So, for signal identification,
is this system accurate? Yes. Is it necessary?
Maybe. And that’s another thing
we’d like to look at trending over time; where there events that
either didn’t get pulled out or got pulled out late? And, what do you think
the most common adverse event reported in this
population was?>>Male Speaker: [inaudible]>>Mary Hardy: What?>>Male Speaker: [inaudible]>>Mary Hardy: Nope!
Trouble swallowing the pill. [laughter] Okay. Now, as we’ve tried to look at taking this stuff
onto analysis. Okay, now we’ve told you
all the ways in which you can find this material,
what you might be looking at, you’re prethinking ahead
about causality. So, basically, you’re not
going to really find– again, I said you’re not
going to find causality. You’re mostly going to be
looking for sentinel events. And these are events which, after you analyze it through
all the data that you have, you can’t find
any other explanation. Either the population is very
young to have a heart attack, or they had a seizure with
no apparent seizure history, no history of trauma, et cetera.
So, that’s a sentinel event: something you can’t explain
that makes you go, “uh-oh.” And then a signal means
that you get a number of these. Now, signals usually
require action, or they often require action, but it’s not like
a research action where you’re then going to go do
a randomized clinical trial to confirm this signal. No, you get
a public health intervention, where someone decides
to send a warning letter, withdraw the material, make
a public service announcement. So, sentinel events
plus leading to signals lead to other kinds
of events. Okay, let me skip this. Now, I’m just going to walk
quickly through the ephedra
example of this. This is an exemplar
so you can see how these different
risk assessments and how the safety assessment
played out. Okay, so first of all, starting with the ephedra
traditional knowledge. How many of you know anything about traditional
Chinese medicine? All right, so you’ll kind of
get the lingo here, but most people are going
to find this like, “what are you talking about?” So, it’s characterized
as an exterior releasing herb recommended for short term
use only. So, I don’t know what
an external releasing herb is if I’m not in traditional
Chinese medicine, but I sure get this part:
short term use only. So, people– already I can tell
you from traditional knowledge, it would be against
traditional knowledge to use this herb
for a long time. Then, here’s the other thing
that it can cause, or can relieve, you know,
and how did they use it? Okay, and if you have
an overdose, they say it’ll take
a couple of hours. It should show up a couple
of hours after ingestion and should resolve about the
same amount of time, roughly. So, that’s from
the traditional knowledge. And then here’s
the pharmacokinetics, and we know what the area
of the curve looks like, when the to Tmax is,
et cetera. So you’ve kind of
got a sense now of within what timeframe of
exposure would we be looking at. So, this is telling us– this is another
one of those things where we show
you millions of things and then tell you
what we actually get down to. So, it turns out
that we got three batches of FDA files from MedWatch. There were 400 of those which
we did a detailed review on; we rejected 87, and these last 333
we did extensive analysis on. Now, because this was
a material that had had a lot
of public interest, there were patient
charts attached. Which sounds benign,
until you realize this patient
probably spent three days or four days
or five days in the ICU. Some of the charts were,
like, hundreds of pages. So, this analysis was very,
very rigorous. We wanted to be shown,
if possible, what was the exact time
the person took the material? Do we have the material
to recover? If there was a death involved,
do we have the autopsy report? Was there evidence in serum
at any point pre or post-death for a measurable level of the
ephedra alkaloids in the blood? Et cetera. So, we were, again, trying
to keep pushing this linkage as tightly as we possibly could. So here’s what we came up with.
Now, this was also– we looked at a whole bunch
of randomized controlled trials as well as adverse events,
so our charge had two parts: one to look at efficacy
and one to look at safety. So, we looked at the safety
from the adverse events reported in
the clinical trials. And as you can see,
we had some adverse events that were reported a lot more
often in the verum group, and the group that took
the ephedra did not. So, this was anxiety,
nervousness, irritability. This could classify as
a lot of different things. Heart palpitations, but then tachycardia
wasn’t really perceived. So, some of this wasn’t
all that consistent. And we could have
expected hypertension based on the pharmacology,
but in fact, even though this was
in a number of trials and had a fair number
of exposed patients, it really wasn’t seen. So, or it may not
have been seen– this may not have been
big enough, because you can see there’s a
difference in the intervention, but it didn’t arrive
at statistical significance. So, it may be then that
if we had another five trials and doubled the amount
of patients, this might become
significant. But, in any case,
some of the things that you expect people
to report– twitchy, irritable, nervous,
insomnia, et cetera– did show up for adverse events. Now, this is the page
from our report, which I can tell you
how to find that– it’s a JAMA article– and if you want
all the gory details, I can tell you how
to find the big report. This is going to go down
by each major adverse event, and we boxed them
into these different boxes. So that we had stand alone. Now, this is going to include
cardiovascular abnormalities that led to death. That got classified as death,
not as a myocardial infarction. So, this is if you survived
your cardiac event, this is if you didn’t. Any kind of cerebrovascular
accidents, seizures, stroke–
I mean, seizures are separate, but any kind of stroke or TIA
or anything like that. Psychiatric symptoms; that we ended up doing
as a separate analysis. But you can see here
that now you go down and you’re going to look
at how many sentinel events did we find? So, we found– and we also
looked at ephedrine in the over the counter reports,
because we thought, just to kind of
put this into context, if we’re going to
look at ephedra the alkaloid is ephedrine. So let’s look and see
what the over the counter– because this is contained in over the counter products
as well. And so you can see
that we only identified two out of the 11 were
sentinel events for ephedra. Now, these other nine
either meant that there was
another causal event, or there was another
causal circumstance, or some other thing
that we thought really could have explained this death.
I’ll give you one example, and we had a big fight
over this. So, there was a young guy who
wanted to make his weight class, and I’m going to–
this is the end of this and then I’ll be able
to take questions– and he put on a sweat suit,
put on a plastic suit, went into the sauna,
took his exercise bike, hadn’t been eating,
took ephedra, and then exercised in the sauna
and died of heat prostration. Well, did the ephedra
cause that? [laughter] Or was it one of the series
of negative adverse events? So, you can see that
that’s what goes into actually arriving
at these sentinel events. If you’re very, very rigorous it can be a very
challenging thing to do. So, with that I’m available
to answer questions and thank you
for your attention. [applause] That’s awfully nice to see,
thank you. Okay, I’m sure somebody–
yes?>>Abby Ershow: Yes, thank you,
interesting talk. This is a thing I was
wondering about yesterday after some of the talks,
and again today. What do you think of registration for almost any dietary
supplement study, NIH sponsored,
industry sponsored–>>Mary Hardy: Yeah,
I think you can–>>Abby Ershow: –backyard
picnics sponsored, would that be useful for–>>Mary Hardy: I think it would
be extraordinarily useful to put some of the more
formalized trials onto that registry, just so that we know where to go
look for data or who to ask, because for this analysis, we actually tried to find
unpublished trials that might have been held
by manufacturers. We, you know, wrote letters,
put in notices, and so just to know where
to go find those trials later, that would be
a terrifically useful event. Now, some of these trials, I think you’re going
to get loaded down with a bunch of stuff
that’s just not even realistic, but I also would like
to see start to include a section for,
“is safety data collected?” Are we going to ask
these questions? Because I think, again
if you want to go back later and look and see, was there a problem
with some of these things, again it would just be great
to know if that was– and it would make people think
about doing it more. So.>>Abby Ershow: Thank you.>>Mary Hardy: You’re welcome. I enjoyed your talk yesterday,
by the way, too. Such a quiet group. You’re really ready for lunch,
is that the problem? [laughter]>>Johanna Dwyer: Not me.>>Mary Hardy: Not you,
excellent! Thanks– Johanna, I can always depend
on you, thank you so much.>>Johanna Dwyer: No,
I have two questions.>>Mary Hardy: Okay.>>Johanna Dwyer: The first is
on ephedra, it seems to me it was
an interesting example–>>Mary Hardy: Uh-huh.>>Johanna Dwyer: –the data
that you’ve shown…>>Mary Hardy: Sure.>>Johanna Dwyer: –and did
such a great report on. But then
the personality factors, and I don’t know
if this is true or not, but one is that a very
high official at FDA’s niece– was it niece, Joe?>>Joseph Betz: Nephew.>>Johanna Dwyer: Nephew.>>Mary Hardy: Oh yeah.>>Johanna Dwyer: He had
an adverse event and I guess he died?
Yeah.>>Mary Hardy: Well,
all my files were de-identified, so I can’t comment about that.>>Johanna Dwyer: Yeah, okay.
And the second one is there was a baseball player, was it–>>Mary Hardy: Yes, yes, yes>>Johanna Dwyer: –over
in Maryland.>>Mary Hardy: Yeah.
Yeah, yeah, yeah.>>Johanna Dwyer: But,
and sometimes the politics works in a direct–>>Mary Hardy: And you know
the– Well, you know
how they found out– how they decided
that he used it? They found it in his locker. Which is true, but I have
a lot of stuff in my locker that I don’t use, so yeah. Those are the adverse events
that publish in the great journal
that we’re all used to: the Boston Globe, the New York
Times, the Washington Post.>>Johanna Dwyer: [laughs]
Democracy dies in darkness.>>Mary Hardy: Yeah, and we
never saw that adverse event, actually.
We never saw that one.>>Johanna Dwyer: Okay.
The second thing is, on TCM–>>Mary Hardy: Uh-huh.
[affirmative]>>Johanna Dwyer: –again,
I’ve heard and I don’t know if it’s true, but hearsay is common
in Washington.>>Mary Hardy: [affirmative] [laughter]>>Johanna Dwyer: –that
traditional Chinese medicine standards for products
sold in China are higher than
they are for export. Is that true?>>Mary Hardy: I have no way
to know that. I know that there is
a commission, or a bureau of traditional
Chinese medicine. And we do know that
when you mess up the food chain or the supplement
chain results– you know, you’re punishment, you
don’t just withdraw your papers. It can be quite harmful. The group that– the minister
that was in charge of the group that passed the melamine milk
and protein, he was actually executed.
Yeah. So, happily,
that doesn’t happen here. Yes, yes?>>Howard Sesso: That was
a nice talk. As I’m guilty of doing
these larger trials and thinking about–>>Mary Hardy: I was thinking
about you!>>Howard Sesso: Well, but when
you’re conducting these types of trials,
no matter how small, medium, large they are though–
I mean, how optimally, when you’re doing trials
of dietary supplements should you be infusing
all the safety, adverse event reporting,
all that too–>>Howard Sesso: Does– like, in a larger scale trial
that we do, it’s harder to dive
into the type of detail that you can with ephedra.>>Mary Hardy: Right. So,
you’re not going to potentially sit down with someone
and have an interview and first let them tell you,
ask an open ended question, “did anything unusual happen
in the last whatever?” And then follow on
to the ones that you, just by knowing the
pharmacology, the physiology, you think might be
more attributable. But like I said, I think
there’s an adverse effect of doing too detailed
of an intervention. You lead your subject, and I’m
a perfect example of that. So, what I would say is,
you ought to make it very, very easy
to have people to report back to you anything
that they didn’t notice, that they noticed
that was unusual. Now, if you have an observant
group, like a physician, then they may tell you,
my urine turned yellow. Well, but, you know they’re
taking a B vitamin supplement, but, so you know some things
are just going to be a wash. I think also, it’s not
unreasonable to include some basic safety data. And even if that’s not you
doing the test, but it may be just
getting permissions signed to get laboratory values
from their physician of record, et cetera.
I would really recommend that. So, yeah, so I do think
just being creative and thinking about this,
I suspect that you could actually end up
with some really valuable data. Again, you’re randomizing
so many people that I think it would be
tremendously useful. Duffy?>>Douglas MacKay: Yes,
Dr. Hardy, thank you. So, you did
a tremendous job sort of depicting the difficulty
of establishing causality when you drill down
and get the patient files, you do all of that.>>Mary Hardy: Right.>>Douglas MacKay: What do you
think of people who are using things
like emergency room visits and data from the Center
for Disease Control to try to make similar
statements about the safety?>>Mary Hardy: Yes.
There’s the DAWN Network, there’s emergency
room visits, there’s a lot
of different ways– the thing is,
there’s a lot of ways to access this kind
of spontaneous occurring events
kind of problems. And at the end of the day, no matter what you’re shown
in that kind of data, it cannot, it cannot
attribute causality. So, we presented this data
that we’d done on ephedra, which was clearly
the longest, deepest– it took us a year,
we had a whole team– was done at
the Southern California Evidence-Based Research Center at the RAND corporation,
for the AHRQ. There’s no one that’s going
to redo that kind of a study. It was enormous.
Even at that, when we presented to the Health
and Human Services group, they were like, “good.
You proved it. We’re done” And you remember there had been
a prior judge’s ruling, which was reversed, because causality was inferred
and which couldn’t be proven, and we said, “yeah, we can
probably prove this. It’s going to be
a case controlled study. It’s going to require
10,000 participants. It’s going to cost
x-millions of dollars, whatever it was going to cost.” And, at the end of the day,
that was not feasible, so a public health decision
was made. Which is a whole different
kettle of fish, and people should be clear
to draw that distinction. And, I have to say,
FDA went way above board to investigate
these serious adverse events. A lot more data was obtained
for those adverse events than you have available for, let’s say, your average
adverse event report. So, you know, I think as long
as we don’t feel like we can do causality from that, we can raise attention,
raise alertness, certainly if any
of these materials are put into a clinical trial,
additional safety data should be collected,
continue pharmacovigilance to the point of finding
a public health risk. And that’s what happened
with ephedra. So.>>Douglas MacKay: Okay,
thank you, Dr. Hardy.>>Mary Hardy: Thank you, guys! [applause]

Leave a Reply

Your email address will not be published. Required fields are marked *