Clinical trial commentary: HPS

1
Clinical trial commentary HPS

• Eric J Topol MD Provost and Chief Academic
  Officer Chairman, Department of Cardiovascular
  Medicine The Cleveland Clinic Foundation Clevela
  nd, Ohio
• Robert M Califf MD Professor of
  Medicine Associate Vice Chancellor for Clinical
  Research Director, Duke Clinical Research
  Institute Duke University Medical
  Center Durham, North Carolina

2
HPS

• Heart Protection Study
• A study carried out on 20 536 patients in the UK
  evaluating use of a fixed dose of simvastatin in
  the prevention of CHD events

3
HPS

• Heart Protection Study
• Patients enrolled40-80 years oldtotal
  cholesterol gt135 mg/dL (gt3.5 mmol/L)At risk of
  CHD due to prior disease
• Randomized tosimvastatin (40 mg) or
  placebovitamin cocktail (600 mg vitamin E, 250
  mg vitamin C, and 20 mg beta-carotene) or
  placebo2x2 factorial design

4
Simvastatin compliance in HPS
Year of follow-up of simvastatin use of simvastatin use
Year of follow-up simvastatin arm placebo arm
Year 5 82 32
Year 6 81 38
Study average 85 18
Collins R et al. HPS trial website
5
Vitamins
"There was absolutely no effect of vitamins,
proving once again that vitamins only enrich the
urine and the people who make them in terms of
coronary disease prevention." Califf
6
Primary endpoints in HPS
Primary endpoint simvastatin (n10 269) placebo (n10 267) p value
All-cause mortality events () 1328 (12.9) 1503 (14.6) lt0.001
Deaths from heart disease and related blood vessel disease () 791 (7.7) 943 (9.2) lt0.002
Collins R et al. AHA 2001
7
Simvastatin major vascular events
Vascular event statin (n10 269) placebo (n10 267) placebo (n10 267)
Total CHD 914 (8.9) 914 (8.9) 1234 (12.0)
Total stroke 456 (4.4) 456 (4.4) 613 (6.0)
Revascularization 926 (9.0) 926 (9.0) 1185 (11.5)
ANY OF ABOVE 2042 (19.9) 2042 (19.9) 2606 (25.4)
Collins R et al. HPS website
8
Long-term treatment
The event curves began together and gradually
diverged over the course of the study "This is a
treatment that should not be given for the short
term, but should be given, probably, for
life." Califf
9
Simvastatin effect by baseline LDL
LDL reduction () LDL reduction () LDL reduction () vascular events vascular events
Baseline LDL (mmol/L) statin(n10 269) statin(n10 269) placebo (n10 267) placebo (n10 267) statin(n10 269) statin(n10 269) placebo (n10 267)
lt3.0(116 mg/dL) lt3.0(116 mg/dL) 1.8 2.7 2.7 602 602 761
3.0-lt3.5 3.0-lt3.5 2.2 3.2 3.2 483 483 655
? 3.5(135 mg/dL) ? 3.5(135 mg/dL) 2.7 3.7 3.7 957 957 1190
All patients All patients 2.3 3.3 3.3 2042 2042 2606
Collins R et al. HPS official site
10
Summary

• Simvastatin showed efficacy for a broad group of
  patients, including many who wouldn't meet
  current treatment guidelines
• Benefit was modest, but present in all subgroups
  divided by age, sex, and baseline LDL level
• No appreciable downside
• Califf

11
Patient safety measures
Blood enzymes (x upper limit of normal) patients showing elevation patients showing elevation
Blood enzymes (x upper limit of normal) simvastatin (n10 269) placebo(n10 267)
Liver enzymes (ALT gt3x ULN) 77 (0.8) 65 (0.6)
Muscle (CK gt10x ULN) 9 (0.09) 5 (0.05)
Collins R et al. HPS official site
12
Prior diseases as entry criteria

• Patients had to have an increased risk of CHD
  death due to prior disease to enter the trial
• MI or other CHD
• occlusive disease of noncoronary arteries
• diabetes mellitus or treated hypertension
• Statins or vitamins not clearly indicated or
  contraindicated according to patient's own
  doctors

13
Vitamins in HPS
Events vitamin (n10 269) placebo (n10 267) p value
All-cause mortality () 1443 (14.1) 1388 (13.5) NS
All vascular mortality () 895 (8.7) 839 (8.2) NS
All nonvascular mortality () 548 (5.3) 549 (5.3) NS
Collins R et al. HPS official site
14
Communication with patients
We must report vitamins as neutral, but "If you
have stuff that costs money, and it has
absolutely no benefit, why on earth would you
take it in the first place?" "How do you convince
people that, just because somebody writes in
Parade magazine that vitamins are good for you,
that they shouldn't be doing this?" Califf
15
Belief in vitamins persists

• Patients take these vitamins and think they are
  helping their cardiovascular health
• "So many trials have put the nails in the coffin
  for vitamins for cardiovascular benefit I don't
  know why it persists."
• Topol

16
Report the negatives of vitamins
" this may push us over the top in terms of
really wanting to convince people to not spend
their money on this but spend it on things that
work." Califf "We're into all this alternative,
complementary, integrative medicine, and if we
could find a vitamin that helps people, we'll
sure recommend it. But if anything, it's sure
going in the wrong direction." Topol
17
LDL levels and mortality
They did not report on mortality alone for the
low baseline LDL group "I sure would like to
know if the survival benefit also showed no
relationship to the baseline LDL." Topol
18
Challenging previous wisdom

• Fascinating results for the low LDL group
• overturns CARE and others that implied a plateau
  of benefit for LDL-lowering
• AFCAPS/TexCAPS showed benefit when patients had
  high CRP and low LDL
• Have we reached the point of not even measuring
  LDL and instead just giving a statin?
• Topol

19
Reserving judgment
We do need to see more detailed data "As you
know, there is nothing that would be more
exciting to someone like me than being able to
say, 'Don't worry about measuring LDL
cholesterol, the whole concept was wrong to begin
with.'" Califf
20
Simvastatin effect by baseline LDL
LDL reduction () LDL reduction () LDL reduction () vascular events vascular events
Baseline LDL (mmol/L) statin(n10 269) statin(n10 269) placebo (n10 267) placebo (n10 267) statin(n10 269) statin(n10 269) placebo (n10 267)
lt3.0(116 mg/dL) lt3.0(116 mg/dL) 1.8 2.7 2.7 602 602 761
3.0-lt3.5 3.0-lt3.5 2.2 3.2 3.2 483 483 655
? 3.5(135 mg/dL) ? 3.5(135 mg/dL) 2.7 3.7 3.7 957 957 1190
All patients All patients 2.3 3.3 3.3 2042 2042 2606
Collins R et al. HPS official site
21
Vascular events by LDL level
Baseline LDL (mg/dL)
Baseline LDL (mg/dL) simvastatin (n10 269) placebo(n10 267)
lt100 285 360
100-129 670 881
130 1087 1365
Collins R et al. HPS trial website
22
LDL as a surrogate

• LDL has held up well for many years as a
  surrogate, now it might be shot down
• "I think the combination of cerivastatin and HPS
  really gives the one-two punch to the concept
  that one can just develop a drug based on
  lowering LDL cholesterol and then really believe
  that you know what its total effects on human
  health are going to be." Califf

23
Run-in period

• Patients were required to tolerate 40 mg of
  simvastatin daily for 30 days before being
  entered into the trial
• may have "cleansed" the data somewhat
• drop-out rate was not reported, but could
  influence interpretation of the trial
• more time could have been spent addressing some
  of the uncertainties

24
Choice of statin

• Which statin should be used at which dose?
• HPS group implied lovastatin comes off-patent and
  could be used cheaply
• Crestor (rosuvastatin) and the other
  "superstatins" offer even more LDL-lowering

25
Statin strategy

• Do we stay with our current strategy?
• we currently start with simvastatin or
  pravastatin because we have the data (and we
  should maybe use a higher dose of simvastatin)
• if LDL doesn't drop enough, we recommend
  atorvastatin this is brought into question by
  HPS
• Califf

26
Practice in light of HPS

• Until we know more about HPS, patients should
  have be informed about their different statin
  choices
• we should use simvastatin 40 mg as reference
  standard
• the unproven statins shouldn't be the first
  choice until they have comparable data
• "Should the clinical world be held hostage to
  unproven theories and marketing strategies?"
  Califf

27
How do statins work?

• Unanswered questions remain about statins
• does lowering LDL more with a given drug actually
  prevent more MIs and strokes?
• pleiotropic effects come into play
• what are the HDL effects?
• "That's the biggest wake-up call of HPS, we don't
  know how the darn statins work." Topol

28
CRP levels and cardiovascular risk
CRP levels (mg/dL) Risk
gt 2.0 High
1.2-1.9 Moderate
lt1.2 Low
Jialal et al. Circulation 2001
29
Lovastatin by LDL/CRP level in AFCAPS/TexCAPS
Treatment event rate lovastatin () event rate placebo () RR (95 CI) p value
All cases ofLDL gt median (n2866) 2.9 5.3 0.53(0.37-0.77) 0.001
LDL lt median CRP gt median (n1428) 2.9 5.1 0.58 (0.34-0.98) 0.04
LDL lt median CRP lt median (n1448) 2.5 2.2 1.08 (0.56-02.08) 0.74
Event rates over 5-year follow-up LDL cutoff
149.1 mg/dL CRP cutoff 0.16 mg/dL
Ridker et al. N Engl J Med 2001
30
Using CRP in the clinic

• How often should you measure CRP?
• needs prospective study
• for now we measure as part of routine blood
  sample in a clinic visit
• if patient is on a statin, CRP is measured when
  they come back for their liver function test
• costs lt8 to run the assay
• Topol

31
Practice in light of HPS

• I don't believe potent LDL lowering means you
  have a better drug, but opinions differ
• work with agents with the longest track record
  (simvastatin, pravastatin)
• save atorvastatin for unmanageable LDL
• in LDL lt100, give a statin if the patient has
  elevated CRP
• Topol

32
Effect of HPS on future trials

• Are trials on statins in the acute phase
  unnecessary now?
• MIRACL
• A to Z
• "I think the window is closing rapidly, and the
  only issue is whether there is something
  different about an unstable plaque in the first
  couple of months."
• Califf

33
Ethics of using placebo

• Can we still justify placebo in cardiovascular
  trials?
• SYMPHONY showed a trend toward detriment for low
  LDL patients
• safety committees need to look at data of
  outstanding trials
• trials like MIRACL seem to have been overshadowed
  by HPS results

34
Use of surrogates in CV medicine

• HPS argues for an end to using surrogates as sole
  criteria for drug development
• still a role for surrogates as a way to winnow
  out therapies
• there is a risk you will miss a treatment because
  it doesn't affect the surrogate of the day
• Califf

35
Cheap and independent trial

• Much was made of HPS as a cheap and independent
  trial
• avoided industry interference
• 30 million for a 20 000 patient trial
• extended follow-up
• Is this a model for future trials?
• Topol

36
Outcome trial

• We need comparative, competitive study with
  clinical outcomes
• surrogates won't tell us enough
• so many beneficial treatments
• beyond the age of placebo in many areas of CV
  medicine
• we should not spend millions on queries about
  data that are irrelevant to the main outcomes
• Califf

37
Independence

• Independence does not mean industry has nothing
  to contribute clinical trial is a shared
  responsibility
• "I would argue strongly that the interpretation
  of the data should be in the hands of people who
  don't work for the company that would stand to
  benefit from the treatment." Califf

38
National research infrastructure

• No national entity designed to support
  cooperative head-to-head clinical trials
• dominant costs subsidized by governments
• controls on industry's access to the data
• this model isn't used enough in the US, where we
  lose trials to industry interference
• Topol

39
A different mechanism

• Once you have developed a scientific principle,
  the rest doesn't happen automatically
• "And I do think that there is increasing
  discussion and insight at the public policy level
  that we're going to have to have a different
  mechanism for outcome-based trials." Califf

40
Topol 2 thumbs up

• "It will be interesting to see in the months and
  years ahead if LDL measurements are abandoned,
  and what the clinical community decides to do
  regarding HPS's findings." Topol

41
Califf 2 thumbs up

• "If we can just talk people who are currently
  taking vitamin E to stop and donate half of the
  money that they were spending to worthy causes,
  we would not only improve the health of the
  population but also contribute to lots of other
  worthwhile endeavors." Califf