Transcript Identifying early signals

David Coulter
IDENTIFYING EARLY SIGNALS
Dar es Salaam
27 Nov 2009
1
Summary of content
 What is a signal?
 Recognising a signal
 What can be achieved by you?
 Clinical assessment of individual events
 Clinical review of collated events
 Principles of signal detection
 Ta
Dar es Salaam
27 Nov 2009
2
Definition 1
A signal refers to ‘reported information on a
possible causal relationship between an
adverse event and a drug, the relationship
being unknown or incompletely documented
previously’.
WHO
Dar es Salaam
27 Nov 2009
3
Definition 2
 In practice it means, a strong suspicion of an
adverse reaction that has not been
recognised previously
Dar es Salaam
27 Nov 2009
4
Recognising a signal
Dar es Salaam
27 Nov 2009
5
Signal identification
 Record
 Collate
 Look!!
Dar es Salaam
27 Nov 2009
6
Recognition of a signal 1
How do we know when events are not
recognised reactions?
 Martindale*
 DrugDex*
 Physicians Desk Reference (PDR).
All available on website of Micromedex
Healthcare Series www.thomsonhc.com
Dar es Salaam
27 Nov 2009
7
Recognition of a signal 2
 You don’t need to do data mining (BCPNN),
or proportional reporting ratios (PRR), or
disproportionality analysis to identify signals
 Careful clinical assessment of your own
events data is the quickest and most
satisfying way.
Dar es Salaam
27 Nov 2009
8
Recognition of a signal 3
 Routine clinical appraisal facilitates
the earliest possible generation of
hypotheses
 Automated signal detection
 good for testing hypotheses
 identifying missed signals
 still needs clinical confirmation
Dar es Salaam
27 Nov 2009
9
Recognition of a signal 4
Clinical review the quickest method
 careful
 informed
 systematic
 standardised
 clinical review
 In your centre
Dar es Salaam
27 Nov 2009
10
Dar es Salaam
27 Nov 2009
11
What can be achieved –by you?
Dar es Salaam
27 Nov 2009
12
What can be achieved?
Example: IMMP -omeprazole
 Hyponatraemia
 Dry mouth
 Taste disturbance
 Interstitial nephritis
 Polydypsia / polyuria
 Polymyositis
Dar es Salaam
27 Nov 2009
13
Omeprazole
 Hepatitis
 Angioedema / urticaria
 Bone marrow depression
 Carcinoid tumour
 Gastric polyps
 Diarrhoea
Dar es Salaam
27 Nov 2009
14
Omeprazole
 Hallucinations
 Amnesia / confusion
 Headache
 Myalgia
 Gynaecomastia / galactorrhoea
Dar es Salaam
27 Nov 2009
15
Omeprazole
 Paraesthesia
 Pruritus
 Rash
 Extrapyramidal symptoms
 Blood dyscrasias
Dar es Salaam
27 Nov 2009
16
Clinical assessment of
individual events
Dar es Salaam
27 Nov 2009
17
COX-2 inhibitors and disturbance of vision
EXAMPLE 1
Dar es Salaam
27 Nov 2009
18
Example 1
 M 78
 Shoulder pain
 Rofecoxib 50 mg once
 Woke next morning with
 no vision right eye
 6/18 left eye
 Recovered next day
Dar es Salaam
27 Nov 2009
19
Example 1
 M 81
 Osteoarthritis knee
 Celecoxib 100mg daily
 Central loss of vision
 Onset after each morning dose, recovering
after a few hours
 No recurrence after withdrawal
Dar es Salaam
27 Nov 2009
20
Example 1
 These 2 case reports can be called the INDEX
CASES
 Contain good information
 close time relationship
 positive dechallenge
 one had rechallenge
Dar es Salaam
27 Nov 2009
21
Example 1
Now we look for information that may
strengthen the signal:
 Other case reports
 WHO database (Vigibase)
 Literature
 Mechanism
Dar es Salaam
27 Nov 2009
22
Example 1
Other reports of eye problems
-blurred vision
Patient
Dose
Onset
Rof
M 58
?
1 week
Cel
F 53
200mg
4 months
Cel
F 59
200mg
1 week
Cel
F71
200
?
27 Nov 2009
Dar es Salaam
23
Example 1
WHO reports
 Celecoxib
 Blindness 12
 Temporary blindness 4
 Vision abnormal 181
 Rofecoxib
 Blindness 22
 Temporary blindness 5
 Vision abnormal 167
Dar es Salaam
27 Nov 2009
24
Example 1
Literature search
 One case report with celecoxib
 Orange spots in both visual fields. (Lund &
Neiman, 2001)
 No reports with rofecoxib
 Visual field defects have been reported rarely
with the traditional NSAIDs
Dar es Salaam
27 Nov 2009
25
Example 1
Mechanism
 Interference with retinal blood flow by
inhibition of prostaglandins and related
substances.
Dar es Salaam
27 Nov 2009
26
Example 1
Conclusion
 Two good index cases
 Several supporting cases
 Supporting cases in WHO database
 Similar reports for related drugs
 A plausible mechanism
 Only one similar report in the literature
 We have a signal!
Dar es Salaam
27 Nov 2009
27
Clinical review of collated
events
Dar es Salaam
27 Nov 2009
28
COX-2 inhibitors and prothrombotic disorders
EXAMPLE 2
Dar es Salaam
27 Nov 2009
29
Demo
 Cluster of events
Dar es Salaam
27 Nov 2009
30
at
o
8
10
3
1
3
0 0
3
0
4
0
0 0
5
6
1
2
1
0
Dar es Salaam
en
ita
l
20
ro
g
51
U
29
Sk
in
0
H
2
Ey
es
5
ae
m
4
H
1
T
4
EN
irc
ul
oc
at
or
rin
y
e/
M
et
ab
ol
ic
C
om
ic
% of Total Incidents
40
lo
gi
ca
l
ep
at
ob
ili
Im
ar
y
m
un
ol
og
ic
al
In
fe
ct
M
io
us
ns
cu
lo
sk
el
et
al
N
eo
pl
as
m
N
s
eu
ro
lo
gi
ca
l
Ps
yc
hi
at
ri
c
R
es
pi
ra
to
ry
En
d
s
en
ta
ry
ut
on
lim
cc
id
en
t
5
A
A
A
45
Profile of Incidents - Celecoxib and Rofecoxib
n=131
n=71
Celecoxib
Rofecoxib
35
30
25
23
15
8
7
9 5
5
4
27 Nov 2009
6
3
4
0
System Organ Class
31
Dar es Salaam
27 Nov 2009
32
Dar es Salaam
27 Nov 2009
33
Prothrombotic events
Summary of findings
 No difference in rates of IHD / stroke between
rofecoxib & celecoxib
 Higher rate of prothrombotic events than
comparators
 Shorter time to onset of death than
comparators
 Differences in death rates due to
prothrombotis events
 Higher rate of cardiac dysrythmias with
celecoxib
Dar es Salaam
27 Nov 2009
34
Principles of signal detection
Dar es Salaam
27 Nov 2009
35
Remember
 Treatment dates -starting date & ending
 Date of onset of event
 Was the patient on the drug when the event
began?
 Calculate onset time
 Effect of dechallenge / rechallenge
Dar es Salaam
27 Nov 2009
36
Signal assessment
Other questions
 Could the problems be caused by a disease?
 The disease being treated
 A co-morbid condition
 Could the problems be caused by another
drug?
 Are the events caused by related drugs?
 Is it relevant or important?
Dar es Salaam
27 Nov 2009
37
Look for non-random features
 Gender
 Age
 Duration to onset
 Survival / life table analysis
Dar es Salaam
27 Nov 2009
38
Non-random features
 Differences in means
 Patients with reaction v patients in cohort
 t-test
 Differences in rates
 RR with CI
 Survival or life table analysis
 Clustering around a certain duration
 Differences between medicines
 Multiple logistic regression
Dar es Salaam
27 Nov 2009
39
Collate reports clinically
 By Clinical Category (CC)
 Then in clinically related groups
 Anatomical functional change
 Clinical sub-group
 Primary event term
 Secondary event term
 Motto:
Sort & see & pursue
Dar es Salaam
27 Nov 2009
40
Identifying early signals





Report your signals to:
your advisory committee &/or regulatory
authority
local health practitioners
the Uppsala Monitoring Centre
local ADR bulletin
medical journal
Dar es Salaam
27 Nov 2009
41
Signal identification
 Record
 Collate
 Look!!
Dar es Salaam
27 Nov 2009
42
Dar es Salaam
27 Nov 2009
43
Merci
beaucoup
Thank You
Dar es Salaam
27 Nov 2009
44