This document discusses detecting cardiac safety signals from drugs. Common drug-related safety issues can be detected in clinical trials by comparing rates to background. Rare but severe issues sometimes appear in trials or are detected by biomarkers or epidemiological studies. Spontaneous increased event rates with a drug usually require a large controlled trial or meta-analysis to interpret. QT prolongation is a safety biomarker and rare torsade de pointes requires alternative risk assessment methods. Diabetes drugs require evaluation of cardiovascular risk through clinical trial design. Acceptable risk bounds are defined to determine if post-marketing studies are needed. Groups like Cardio-Oncology and the Cardiac Safety Research Consortium conduct research on these issues.

1. Special cardiac safety
concerns
Shari L. Targum, MD, MPH, FACC
Medical Officer
U.S. Food and Drug Administration

2. 2
Detecting cardiac safety signals
– Common, drug-related: can detect in placebo-controlled,
clinical trials, compare to background rate
– Rare, severe, drug-related: sometimes detected in clinical trials (e.g.,
Stevens-Johnson) or via
• risk biomarkers (e.g., QT prolongation) or
• epidemiologic studies (e.g., case-control)
– Spontaneous events ↑ rate with drug: single event usually not
interpretable; detect via
• large enough controlled trial
• compare to background rate
• epidemiologic study (large hazard ratio)

3. 3
QT prolongation and risk of arrhythmia

4. 4

5. 5
QT interval
• Variable-- heart rate, autonomic tone, time of
day
• Can be prolonged due to:
– heart disease (e.g., congestive heart failure)
– electrolyte abnormalities (e.g., hypokalemia)
– drugs (e.g., quinidine).

6. 6
Torsade de pointes (TdP)
Rare, but life-threatening. Might not be detected
in a drug development program.

7. 7
Background
–Drug withdrawals due to QT risk for non-
antiarrhythmic drugs (e.g., cisapride,
terfenadine)
–How to evaluate TdP risk in drug
development (e.g., prior to marketing)?

8. 8
QT as a safety biomarker:
• Era of the “Thorough QT” (TQT) study
• Threshold for potential clinical importance set
very low (10 msec)
• “Negative study” → routine phase 3 monitoring
• Failure to rule out 10 msec → heightened
phase 3 monitoring

9. 9
QT Study Characteristics
• Characterize QT effects of the drug under near
“worst case” scenario
– Exposure at supratherapeutic concentrations
– ECG sampling at peak concentrations
(drug/metabolites)
– Sufficient duration of dosing/sampling to
characterize effects

10. 10
Some concerns about QT studies
• TQT studies difficult and expensive
• QTc relationship to risk (arrhythmia) not
constant
• Unknown public health consequences of
compounds removed from pharmaceutical
pipeline
• Interest in alternative approaches to assess
proarrhythmic risk.

11. 11
SCIPA (Comprehensive In Vitro
Proarrhythmia Assay) Initiative
1. in vitro drug effects, multiple cardiac
channels + in silico reconstruction of electrical
effects;
2. confirmation using human stem cell-derived
cardiomyocytes.
Undergoing validation at this time.

12. 12
Drug-induced Valvulopathy

13. 13
Weight loss and “Fen-phen”…

14. 14
Appetite suppressants
• Fenfluramine (1973): approved for short-term
use
– racemic mixture*- increased serotonin, associated
with depression
• Dexfenfluramine (1996)* thought to be “safer”
• Fen-Phen: never approved, widely used off-
label for long-term management
*withdrawn in 1997

15. Case-control study in Europe: odds ratio 23.1
associated with use > 3 months.

16. Fenfluramine and dexfenfluramine voluntarily withdrawn on Sept. 15, 1997
24 women, no prior heart disease, mean treatment duration 11 months.

17. Source: Bhattacharyya et. al. Lancet 2009; 374: 577-85

18. 18
Diabetes drugs and cardiovascular risk

19. 19
Cardiovascular risk and diabetes drugs
• Diabetes drugs approved based on glycemic
control (hemoglobin A1c)
• Diabetics have increased cardiovascular risk
• Concerns that some medications increase
cardiovascular risk (and little information)
• Need to show that treatment doesn’t result in
unacceptable risk (e.g., non-inferiority)

20. 20
Cardiovascular (CV) risk and diabetes drugs
• Guidance evaluating cardiovascular risk in new
antidiabetic therapies to treat type 2 diabetes
(2008)
– Design Phase 2/3 trials to allow meta-analysis
– Blinded CV endpoint adjudication committee
– Include higher risk patients (e.g., elderly, renal
impairment)
– Prespecified upper bound
• May need adequately powered cardiovascular
outcome study.

21. 21
Defining acceptable cardiac risk and
diabetes drugs
Upper bound of 95% confidence interval
for risk ratio or hazard ratio
Conclusion
>1.8 Inadequate to support approval
>1.3 but <1.8a Postmarketing cardiovascular trial(s)
needed to show definitively <1.3
<1.3a Postmarketing cardiovascular trial(s)
generally not necessary
Acceptable and unacceptable cardiovascular risk
awith a reassuring point estimate

22. Detecting cardiac safety signals
– Common, drug-related: detect in placebo-
controlled, clinical trials of appropriate
duration
– Rare, severe, drug-related: detect via
• risk biomarkers (e.g., QT prolongation)
• epidemiologic studies (e.g., valvulopathy)
– Spontaneous events ↑ rate with drug: single
event usually not interpretable; detect via
• large enough controlled trial
• Meta-analysis (e.g., diabetes drugs)

23. 23
Cardiac Safety-related Groups

24. 24
Cardio-Oncology
• Several oncology drugs associated with cardiac
toxicity (e.g., anthracyclines, trastuzumab,
tyrosine kinase inhibitors)
• Interest in assessing and mitigating
drug/radiation-induced cardiovascular risk
• New interest group (American College of
Cardiology), journal (Cardio-Oncology),recent
FDA public workshop (22 September 2016)

25. 25
Cardiac Safety Research Consortium
• Launched in 2006 through an FDA Critical Path
Initiative Memorandum of Understanding with
Duke University to support research into the
evaluation of cardiac safety of medical products.
• Industry-academia-government effort
• Think tanks, research projects, publications
• Further information: cardiac-safety.org