Choice of Thrombolytic
                        ======================

þ Cost:
 - tPA:  $2000-2700
 - streptokinase:  $200

þ For the elderly:
 - less intracranial bleeds in patients over 75 yo with strepto

þ Heparin
 -should only be given concomittantly with tPA, not with Strepto. Strepto
  itself causes an elevation of the PTT so heparin should not be given until
  six hours after the completion of the strepto infusion or the PTT returns
  to subtherapeutic levels, whichever occurs first.

þ Efficacy:

On Tue, 5 Nov 1996, David B. Toth wrote:

> At 08:41 AM 11/4/96 -0400, John Schoffstall wrote:
>
> >     The issue arose when we had something like six mixed/unused TPA's in
> >a month. All were refunded, but the pharmacy got nervous. Actually, I
> >don't think all were from the ED.
> >
>
> In all seriousness, don't you think you're using too much TPA ???
> I use TPA only about 5% of the time as compared to Streptokinase. It is
> really hard to justify using it more than that, based on cost. If you feel
> justified using it based on very slim stats for 1 type of MI, then what
> about the same slim stats re increased side-effects?

      Well, it only took me about month to reply to this. <sigh> I think
it took me about a week after getting back from vacation just to deal with
the accumulated mail. I wish there were killfiles for snail mail.
Spectrum? *plonk* Special low rates for Newsweek? *plonk* The dean didn't
get his faculty survey returned, and wants *me* to fill out thirty
scannable bubble forms with a No. 2 pencil? *plonk* I have a little list,
they never will be missed. Or if you could just turn off your physical
mailbox during vacation, the way you can turn off emed-l...

     Okay, okay, to business. I'm going to reply not only to your post,
but to a couple others that were made in reply, or in related threads, so
please bear with me. Apologia pro thrombolytica mea. Do I think I'm using
too much tPA? No. I don't.

     Let's look at the 'slim stats'. For 30 day mortality, Gusto found
that there was a 1% benefit in overall mortality, 7.3% vs 6.3% between tPA
and both streptokinase strategies, p < .001. There was still a significant
benefit when compared to either strepto strategy individually.[1] How big is
this benefit? 1%? Or is it 14%, i.e., you have a 14% (1/7.3) less chance
of dying of your MI (95% C.I. 5.9% to 21.3%) if you are given tPA rather
than strepto?

     It's 14%. The assessment of benefit must be related to the population
at risk.[2] 93% are going to live if they get *any* thrombolytic regimen.
The ones at risk are the 7% who are going to die if they get strepto. You
can save 14% of those fated to die by giving tPA.

     Is 14% 'slim'? Not for me. Not when we're talking about an endpoint
of death. I wouldn't call it small if it were *my* life at risk.

     Bear in mind that a 7% mortality rate is still pretty low, and that
further gains are going to be difficult. The easy stuff has been done.
Future improvements in MI mortality may only come in absolute increments of
1% or so, and will need large studies to detect reliably.

     How about 'the same slim stats re increased side-effects'? No. I
think you're making the error of counting the side-effects twice: once
because including patients with fatal hemorrhagic stroke in the deaths
makes the efficacy seem 'slimmer', and including them again as
'side-effects'. Including side-effects, tPA is still better than strepto.

     How about '1 type of MI'? No. The figures above are for all MI's.
Sub-group analysis shows advantage for both anterior (10.5% mortality for
strepto, vs 8.6% for tPA) and inferior MI's (5.3 vs 4.7). The CI for the
95% odds ratio for anterior MI's does not cross unity, the CI for the 95%
odds ratio for inferior MI's does cross unity, i.e., does not meet
criteria for an alpha error of .05.

     But, as has been said, no proof of a difference is not proof of no
difference. If you slice and dice any group into subgroups, you're going
to lose power to prove anything. Is tPA less effective in inferior MI's?
Quite possibly, probably because they have a lower mortality rate to begin
with. But is it no better than strepto?  Please take a look at the actual
CI's.[3] The trend is clear. I would have very hard time trying, with a
straight face, to claim from this data that there is no evidence that tPA
is better than strepto.

     Likewise with age. The finer you slice the data, the less
significance you get, but the trends are clear, and they are all in the
same direction: favoring tPA. The benefit in patients > 75, for example,
was the same as in the overall study population: it probably didn't reach
significance simply because the numbers were smaller.[4] With the
exception of hemorrhagic stroke, complication rates, including allergy and
anaphyllaxis, CHF, cardiogenic shock, VT, VF, and so on, all favor tPA
over strepto, and the difference is < .05 in all cases.[3]

     If the data were mixed, with some groups benefiting more from tPA
and some from strepto, some complications more frequent with tPA, some
with strepto, I would buy that perhaps there was really no difference.
But the data aren't mixed. They all favor tPA -- again, excepting
hemorrhagic stroke. In some groups, there are large enough numbers to
prove a difference at the .05 alpha error level, in some groups there
aren't, but in all groups the trend is unmistakable.

     You say you find it "really hard to justify using it more than [5% of
the time], based on cost." How much is the cost? Well, given the
difference in price between tPA and strepto, it works out to about
US$200,000 for every life saved if tPA is used rather than strepto.
That's a lot. But the average survivor will live 9-11 years, making
incremental cost per year of life saved about $20,000. (If we get into
'quality adjusted year of life' the figure will vary.) When we ask if we
can justify this, we should look at whether we, as a society, believe in
other interventions that cost this much, per year of life saved. As other
discussants have recently noted, this figure is not out of line with what
society commonly spends: Robert Wears notes that a study of 293 medical
interventions found a cost of $17,000 per year of life saved, and that
environmental risk reduction measures -- which generate almost universal
public approval, probably because none of us feels that he is the one
paying for them -- may cost $350,000 per year of life saved. The GUSTO
investigators compare tPA to antihypertensive therapies. These drugs are
much cheaper than tPA, but they also must be used more frequently and in a
larger number people to save a life, so the net cost of a 'quality
adjusted year of life' saved is $33,000 for tPA vs $20,000 for treating
severe hypertension. For renal failure, the cost is $35,000.[5] Another
group [6] found $27,000-30,000 the cost per year of quality adjusted life
(QALY) saved by using tPA rather than strepto. This group did a sensitivity
analysis, varying the theoretical benefit of tPA vs strepto, and the
cost. Even if you cut the mortality benefit of tPA to a third of what was
found in GUSTO, you still get a cost per QALY of $109,400, close to
the upper limit of what is generally felt to be tolerable to society. If
the price of tPA were lower, the cost per QALY is only 9,000, but remains
less than $50,000 as long as the price of tPA is < $4,400 per dose.
Interesting article. Worth pulling and reading.

     So, that's my take on tPA vs strepto. I believe, based on what I've
said above, is that the what has been commonly said in this thread about
tPA not being effective in patients > 75, or < 55, or inferior MI's, or
whatnot, is nonsense, and based on a failure to look closely at the data
and deal honestly with its implications. tPA is cost effective, compared
to what society commonly spends on medical interventions.

     But I can't let this thread get away without bashing Genentech at
least once. This whole discussion would never have to happen, and
well-meaning physicians would never be lured by illusory cost savings into
using an inferior drug *if* Genentech would price its product reasonably.
A cynic would say that Genentech picked the price so as to hit that $/QAYL
target. At least, I'm sure they arrived at the price based on some pretty
heavy duty computer runs to maximize worldwide profits, considering the
price of strepto, the cost of advertising, the length of patents, and so
on, and that the price has little to do with the actual cost to develop
and manufacture the drug.

     I don't deny their right to price their drug any way they choose. But
I consider it very badly behaved of them to price it the way they did.
They have garnered much ill-will in doing so.

     Three final thoughts:

     1) It's absurd, in a way, to spend so much time on this issue of tPA
vs strepto. Energy would be better spent in trying to find way of getting
patients with chest pain to come to the ED sooner.  Alas, those efforts
have largely been failures. It's easier to change a drug than change human
nature.

     2) I really, really dread going through this all again with
all the pricey CVA and gram negative sepsis drugs in the pipeline.

     3) Apologies to everyone whose server I crashed with the length of
this post. This one post, to 900+ emed-l recipients, took up over 9 megs
of Internet bandwidth.


[1] Lee KL, et al: Holding Gusto up to the light. Ann Int Med
1994;120:876-881.

[2] Sheps MC: Shall we count the living or the dead? N Engl J Med
1958;259:1210-1214. (An old article, but still great. Pull it. Read it.)

[3] Holmes DR, et al: Lessons we have learned from the GUSTO trial. J Am
Coll Cardiol 1995;25[Supplement]:10S-17S].

[4] Topol EJ, et al: letter. N Engl J Med 1994;330:505-506.

[5] Mark DB, et al: Cost effectiveness of thrombolytic therapy with
tissue plasminogen activator as compared with streptokinase fro acute
myocardial infarction. N Engl J Med 1995;332:1418-1424.

[6] Kalish SC, et al: A cost-effectiveness model of thrombolytic therapy
for acute myocardial infarction. J Gen Int Med 1995;10:321-330.

John Schoffstall, M.D., aka schoffstall@allegheny.edu
Department of Emergency Medicine
Allegheny University of the Health Sciences, Philadelphia, PA
"When all think alike, no one thinks very much." -- Walter Lippman
--------------------------------------
On Thu, 12 Dec 1996, Dan Liao, MD wrote:

> There was a lengthy discussion about tpa vs. strepto recently.  Several
> people mentioned that tpa showed no advantage over strepto for MI's over
> three hours old.  Thus some ED's use strepto for MI's over three hours old
> and tpa for MI's less than three hours old.   I'd like to know the
references
> for this information. You may email me personally so as not to congest the
> EMED list with this information. Thank you for your help! Dan Liao, M.D.
>   (DnLiao@aol.com)

The reference is from GUSTO (1).  Patients treated more than 4 hours after
symptom onset actually had a non-significant trend toward improved outcome
with streptokinase. Between 0 and 2 hours there was a relative
advantage of tPA over streptokinase of 20%. By  2 to 4 hours it was
18%.  At 4 to 6 hours it had shrunk to only 4% and beyond 6 hours
there was actually a 25% advantage of streptokinase over tPA.  When
the data for times greater than 4 hours are pooled streptokinase still
has a (nonsignificant) advantage. Thus the contention that had been made
in a recent post claiming that ALL subgroups showed improved outcome with
tPA (whether statistically significant or not) is incorrect.


This in spite of the fact that GUSTO was designed in such a way that it
'favored' tPA by virtue of 1) comparing tPA to 2 inferior streptokinase
regimens (it had aleady been established that the addition of heparin to
streptokinase added nothing except increased ICH),  2) having a higher
rate of salvage CABG in the tPA group in spite of comparable degrees of
CHF, pulmonary edema and shock and 3) demonstrating that the advantage of
tPA to streptokinase was limited to those patients who were studied in the
US where the rates of bypass and PTCA were considerably higher.


H. Louzon MD


(1) NEJM 1993;29:673-682