Effect of Rosiglitazone on the Risk of Myocardial Infarction and Death From Cardiovascular Cause Nis

1
Effect of Rosiglitazone on the Risk of Myocardial
Infarction and Death From Cardiovascular
CauseNissen, S. E. and Wolski, K. Effect of
Rosiglitazone on the Risk of Myocardial
Infarction and Death from Cardiovascular Causes.
N Engl J Med. 2007 May 21.

• ACCU Conference
• 23 May 2007

2
Chronic Illness Litany
3
Chronic Illness Care Hallmarks

• Patient asymptomatic
• Patient is experienced, free-ranging and
  principal caretaker
• Therapy occurs in unstructured environment
• Symptom reversal unusual not principal goal
• Creating non-events
• Paradox Treating asymptomatic conditions may
  cause symptoms
• Therapy may undermine confidence

4
Perils of Polypharmacy
5
Pharmacotherapeutic Pitfalls The Trauma
Literature

• The Text Don Quixote de La Mancha, Cervantes
• Design Case Study/Series
• Methods/Context A spontaneous brawl between Don
  Quixote and Sancho Paz, and a prostitute and
  innkeeper (procurer)
• The Intervention Healing wounds caused by
  batons, oil lamps and other objects
• Pharmaceutical A boiled reduction of rosemary,
  oil, salt, and wine (referred to as the the
  balm or balsam in some translations

6
The Cure Don Quixote

• Don Quixote wanted to test the virtue of the
  precious balmAs soon as he finished drinking it,
  he began to vomit until nothing was left in his
  stomach, and with the nausea and agitation of
  vomiting, he broke into a copious sweat, for
  which reason he ordered them to wrap him up well
  and leave him aloneHe slept for more than three
  hours, and when he woke his body felt much
  relieved and so much better after his beating
  that he considered himself curedand that with
  this remedy he could from now on, and with no
  fear whatsoever, engage in any combat, battle or
  contest no matter how perilous it may be.

7
The Cur(s)e Sancho Paz

• Sancho Paz, who also deemed the improvement in
  his master a miracle picked up the pot, and with
  a good amount of trust and even greater optimism,
  he gulped the potion down thirstilyIt seems,
  however, that poor Sanchos stomach was not as
  delicate as his masters and so before he
  vomited, he endured so much nausea and felt so
  sick to his stomach, and sweated so much and felt
  so faint, that he really and truly thought it was
  his final hour

8
The Curse (cont.)

• At this point the concoction took effect, and
  the poor squire began to erupt from both
  channels, and with so much force that the reed
  mat on which he lay, and the canvas blanket that
  covered him, could not be used againHe cursed
  the balm and the villain who had given it to him

9
400 Years Later
10
New Dx Diabetes
11
(No Transcript)
12
(No Transcript)
13
The Smoking Gun
14
Rates of Myocardial Infarction and Death from
Cardiovascular Causes
Nissen S and Wolski K. N Engl J Med
200710.1056/NEJMoa072761
15
Risk of Myocardial Infarction and Death from
Cardiovascular Causes for Patients Receiving
Rosiglitazone versus Several Comparator Drugs
Nissen S and Wolski K. N Engl J Med
200710.1056/NEJMoa072761
16
Nature of risk

• Risk seen early (trial duration 24 to 52 weeks)
• Benefits take years to register from statins and
  anti-hypertensives
• Increased risk compared to other DM drugs (Table
  5)

17
Mechanisms of Increased Risk

• Adverse lipid profile effects
• Fluid retention, weight gain
• Provoke CHF, and increase in cardiac workload and
  O2 demand
• Reduction in HgB
• Cardiac toxicity seen in other TZD/PPARs whose
  development scuttled by toxicity (muraglitazar)

TZDs are agonists for peroxisome-proliferatorac
tivated receptor (PPAR)
18
A Class Effect?

• What about pioglitazone (Actos)?
• No toxicity CV AEs seen in clinical trials
• More favorable effects on lipids
• Is glycemic control alone a reliable surrogate or
  intermediate outcome for diabetes management?
  (predicts micro- but not macrovascular
  improvements)

19
Limitations

• Inaccessibility of source data prevented
  construction of composite outcome of MI or death
  from CV causes
• MI not consistently defined
• No standard outcome validation
• Small trials of short duration
• Unable to conduct time-to-event or dose response
  analyses

20
New Drugs Are Less Risky

• True in a Crude Statistical Sense
• No track record (100 years of ASA experience)
  and few events accumulated
• Uncomplicated patients
• Unadjusted for comorbidity
• Less polypharmacy in study patients
• No long-term follow up

21
AE Information Deficit

• Inadequate post-marketing surveillance
• Primary focus on narrowly defined efficacy
  measures in RCTs
• Patient factors RCTs study patients with few
  comorbidities with fewer meds
• Beta-error Insufficient power to detect AEs
• Researcher conflict of interest
• Scientific misconduct

22
FDA Regulatory Failure

• Design of Phase 4 (post-marketing trials)
  specified by pharma
• Only 25 of Phase 4 trials completed between 1998
  and 2003
• Reliance on surrogate outcomes (A1c) rather than
  clinical outcomes
• Need for life-cycle approach to balance
  benefits and risks of medication

23
Improving AE Information

• More thought to AEs a priori in trial design
• Conduct RCTs in representative patients
• Combine AE rates from multiple trials
• Standardize definitions of harm (rheum, oncology)
• Insist on appropriate interpretation of
  non-significant P-values with AEs (Beta error)