Eugene Braunwald MD

1
Thrombolysis and the elderly

• Eugene Braunwald MD
• Professor of Medicine
• Harvard University
• Robert Califf MD
• Professor of Cardiology
• Duke University
• Eric Peterson MD
• Associate Professor of Cardiology
• Duke University

2
Thrombolysis and the elderly

• Lack of benefit for thrombolysis in
  patients gt 75 years
• This retrospective cohort was taken from a
  database compiled by the Health Care Financing
  Administration (HCFA) for the Cooperative
  Cardiovascular Project (CCP) detailing a
  nationwide sample of elderly patients with MI.
• From 210 996 original patients, a cohort of 5191
  patients aged 65-75 and 2673 patients aged 76-86
  was obtained.
• Among exclusion criteria were the following
  direct admission to hospitals with on-site
  angioplasty capabilities, inappropriate ECG
  criteria on presentation, and contraindications
  to thrombolysis.

Thiemann DR, et al. Circulation 20001012239-2246
3
An increase in mortality?
Observed 30-day mortality
Age 65-75 Age 65-75 Age 76-86 Age 76-86
without thrombolysis (n1330) with thrombolysis (n3861) without thrombolysis (n1066) with thrombolysis (n1607)
Crude 9.8 6.8 15.4 18.0
Predicted 9.6 7.8 15.3 13.6
predicted in the absence of thrombolysis predicted in the absence of thrombolysis predicted in the absence of thrombolysis predicted in the absence of thrombolysis predicted in the absence of thrombolysis
Thiemann DR, et al. Circulation 20001012239-2246
4
Multivariate adjustment
Adjusted hazard ratios for 30-day mortality
after receiving thrombolysis
Age 65-75 (n5113) Age 76-86 (n2634)
First model 0.76 (0.61-0.95) 1.29 (1.06-1.58)
Second model 0.88 (0.69-1.12) 1.38 (1.12-1.71)
second, or comprehensive model, adjusts for multiple variables in addition to those present in the first model second, or comprehensive model, adjusts for multiple variables in addition to those present in the first model second, or comprehensive model, adjusts for multiple variables in addition to those present in the first model
Thiemann DR, et al. Circulation 20001012239-2246
5
Increasing hazard with age
Hazard ratios for thrombolytic therapy increase
with increasing age

• Age Hazard ratio 95 CI
• 65 0.60 0.44-0.82
• 74.3 1.00
• 80 1.36 1.13-1.64

Thiemann DR, et al. Circulation 20001012239-2246
6
Thrombolysis and the elderly
Observational studies

• This study is observational in nature, and not a
  substitute for a randomized, controlled, clinical
  trial.
• In observational studies it is difficult to
  adjust for all possible variables and to control
  bias.
• Observational studies are, when well designed,
  able to raise and highlight important questions.
• The safety and efficacy of thrombolytics in an
  older population (76-86 years old) is indeed an
  important question.

Thiemann DR, et al. Circulation 20001012239-2246
7
Thrombolysis and the elderly
Observation vs randomization

• Real-world experiences (observational studies)
  often tell us things that a randomized clinical
  trial cannot.
• However, real world experiences can't compare
  treatments in high risk groups of patients where
  treatment selection is based on major prognostic
  factors that are difficult to quantify.
• One concern is that such observational studies
  may impact clinical practice beyond
  reasonableness.
• In the world's compilation of placebo vs active
  treatment there is a trend toward benefit for
  those gt 75 (approximately 2000 patients ).

8
Thrombolysis and the elderly
Need for additional studies?

• An observational study is notable because of the
  lack of information in gt 75 age group in prior
  randomized trials.
• Although a trend toward benefit exists in those gt
  75 under active treatment (world data), this
  trend is a lot less than that seen in younger age
  groups.
• Earlier thrombolytic trials used different agents
  than those studied today.
• In performing a randomized controlled trial,
  sufficient numbers of patients of age greater
  than 75 need to be enrolled.

9
The role of observation
If you think about how we accumulate medical
evidence, we first make an observationfrom that
observation you set up a design to study a large
number of patients prospectively. Sometimes the
first observation that youve made bears out in a
trial, and at other times it does not. Dr
Eugene Braunwald Professor of Medicine Harvard
University
10
The role of databases

• Approach databases in a manner similar to that
  for the unusual or interesting patient. You may
  not be able to draw many conclusions, but you may
  glean some hints.

eg, analysis of the database for patients with
rheumatoid arthritis revealed that patients on
ASA had a lower incidence of death from MI large
number of ASA trials in MI prevention and
treatment ASA accepted in MI management
11
The ideal patient

• Placebo controlled trials can no longer be
  performed with patients who are otherwise ideal
  candidates for thrombolytic therapy.

Ideal thrombolysis candidate
50 years old marked ST elevation 45 min crushing chest pain no contraindications to lysis
12
Thrombolysis and the elderly
CCP database

• Characteristics of the CCP database
• comorbidity
• some functional status data
• dementia variables
• Variables often provided as yes/no.
• Subtleties are not available in any database,
  hence unmeasured confounding is possible.

13
SHOCK trial

Methodology

• Patients with MI and left ventricular failure
  were randomized to emergency revascularization
  (n152) or initial medical stabilization (n150).
• Revascularization was defined as either coronary
  artery bypass grafting or coronary angioplasty.
• The primary endpoint was all-cause 30-day
  mortality.
• All-cause 6-month mortality and additional
  subgroup analyses were investigated.
• Mean age of the patients was 66 10 years and
  32 were women.

Hochman JS, et al. New Engl J Med 1999341625-634
14
SHOCK trial

Results
Overall mortality () Overall mortality () Overall mortality () Overall mortality ()
Revascularization n152 Medical therapy n150 Relative risk (95 CI)
30-day 46.7 56.0 0.83 (0.67-1.04)
6-month 50.3 63.1 0.80 (0.65-0.98)
Mortality in age group gt 75 () Mortality in age group gt 75 () Mortality in age group gt 75 () Mortality in age group gt 75 ()
Revascularization n24 Medical therapy n32 p-value
30-day 75.0 53.1 1.41 (0.95-2.11)
6-month 79.2 56.3 1.41 (0.97-2.03)
Hochman JS, et al. New Engl J Med 1999341625-634
15
GUSTO-I trial data

Outcomes in the elderly

• Analysis of the GUSTO-I trial data revealed that
  30-day mortality increased markedly with age
• lt65 n24 708 3.0
• 65 to 74 n11 201 9.5 30-day
• 75 to 85 n4625 19.6 mortality
• gt85 n412 30.3
• Combined death or disabling stroke appeared to
  occur less often with TPA than with streptokinase
  in all but the oldest patients who demonstrated a
  weak trend suggesting a lower incidence with
  streptokinase and s/c heparin odds ratio 1.13
  95 CI 0.6, 2.1.

White HD, et al. Circulation 1996941826-1833
16
Clinical trials in the elderly

Importance in health policy

• The need for ongoing clinical trials in the
  elderly is apparent, especially given the aging
  of the baby-boomer population.
• Although current evidence supports the idea of
  giving thrombolysis, specific trials in this age
  group become ethical and very important.
• The role of percutaneous intervention is also at
  issue.
• Patients gt 75 make up 1/3 of all patients with
  MI, but over 1/2 of all deaths.

17
Clinical trials in the elderly

Early hazard with thrombolytics

• 30-day mortality data in a thrombolytic trial of
  the elderly may reflect an early hazard due to an
  increase in both intracranial hemorrhage and
  cardiac rupture.
• Benefit might therefore not be seen for 6-12
  months.

18
Clinical trials in the elderly

General implications

• The establishment of randomized controlled trials
  in thrombolysis might aid in informing the
  medical community that the elderly are being
  undertreated.
• Practicing physicians should STOP undertreating
  the elderly, and SHOULD continue to look for
  opportunities to participate in clinical trials.