Response to treatment in ADHD: Prediction and Compliance

1
Response to treatment in ADHD Prediction and
Compliance

• Eric Taylor

Kings College London Institute of Psychiatry
South London Maudsley NHS Trust
2
A complex disorder, multiply caused
3
Goals of treatment

• Reduce core symptoms
• Impulsiveness, inattentiveness, restlessness
• Alleviate comorbid problems
• Aggression-anger, anxiety, emotional lability,
  autism spectrum, Tourette
• Promote realistic learning goals
• Reduce rejection by others
• Understanding of disability, valuing of self and
  of authority

4
WHY PREDICT RESPONSE?

• (A treatment trial will often give a clearer
  answer than even a significant prediction)
• To understand the treatment
• To influence choice of treatment
• - therapy helpful only for a small minority
• - hazardous or expensive therapy
• - many therapies available
• To influence treatment regime
• - dynamics, kinetics, adherence

5
Value of a predictive marker(Assuming 70
response rate c. 80 sensitivity specificity)

• Improved Treated Utility
• unnecessarily (1 v. -1)
• 100
• 70 30 40
• No pretest
• 100 Positive 60 48 12 36
• Pretest Negative 40 0 0

1 v -3
-20
12
So a marker becomes useful if treatment is
hazardous or very costly (or if there are many
possible treatments and response is slow)
6
Future of pharmacotherapyiInternational
licences now being sought

• Guanfacine
• Dexamfetamine complex
• Modafinil
• Focalin
• Risperidone for irritability

Coming later percutaneous delivery nicotine
GABA analogues AMPAkines CREBS?
7
Predicting outcome

• Single dose
• Clinical profile
• Subtypes
• Comorbidity
• Pharmacogenomics
• Brain function
• Compliance

8
UNDERSTANDING THE TREATMENT

• Measure the marker at baseline and outcome
• Correlate change in marker with clinical outcome
• Avoid confounders
• - regression to mean, recovery, placebo,
  fluctuations
• Compare with placebo
• - crossover
• - regression
• Fixed v variable dose absolute or relative
  outcome

9
REVIEW OF REPLICATED PREDICTORS

• Psychophysiology high or low skin
  conductance
• normal or abnormal EEG
• high or low heart rate
• Neurology presence of soft signs
• Familial good management
• Age younger or older
• Clinical high severity (IQ
  inconsistent) poor attention
• Barkley (1976) Journal of Abnormal Child
  Psychology

10
PREDICTING (DRUG PLACEBO) CHANGE

• Responder Non-responder Iambda
• N22 N16
• Age (months) 95 111
  .85
• IQ 90 98 .84
• PACS/Hyperactivity 1.2
  0.4 .75
• Attention -1.5 0.6
  .72
• Clumsiness 14 8
  .70
• Significant predictors in crossover double-blind
  R.C.T in boys with disruptive behaviour Taylor
  et al (1987) Psychological Medicine, 17, 121

11
What predicts good response?

• Taylor et al Buitelaar et al
• High severity Low severity
• Low IQ High IQ
• Young age Young age
• Low anxiety Low anxiety

12
Differences between studies

• Taylor et el Buitelaar et al
• Disruptive behaviour ADHD
• Optimal dose Fixed low dose
• Marked improvement Normalised
• (improved was not predicted)
• Crossover placebo Regression subtraction

13
Measuring outcome normalisation v. effect size
                                               
               
Unmedicated ADHD
General population
Medicated ADHD
14
(No Transcript)
15
(No Transcript)
16
CLUSTERING IN BOYS WITH DISRUPTIVE BEHAVIOUR

• Cluster Medication response
• Hyperkinetic 73
• Conduct 14
• Anxious/depressed 8
• Marked improvement in drug and not placebo
  N38 in crossover double- blind R.C.T.
• Taylor et al (1986) British Journal of
  Psychiatry, 149, 760-777

17
Month
0
36
24
14
10-m Follow-up Phase
22-m Follow-up Phase
14-m Treatment Phase
R A N D O M A S S I G N M E N T
MedMgt 144 Subjects
Recruitment Screening Diagnosis
Beh 144 Subjects
579 Subjects 7 to 9 yrs old ADHD-Combined
Comb 145 Subjects
CC 146 Subjects
Early Treatment (3 m)
Mid- Treatment (9 m)
End of Treatment (14 m)
First Follow-up (24 m)
Second Follow-up (36 m)
Baseline
Pre-Baseline
Observation 2 LNCG Group
Observation 1 LNCG Group
Assessment Points
18
Comparing TherapiesConclusions from MTA Study

• Medication is more powerful than behavioural
  treatment at 14 months
• Research treatment better than routine
• Many advantages in adding medicationto
  behavioural treatment few in adding behavioural
  treatment to medication

19
Subtyping

• ANXIETY / DEPRESSION

IMP 1/4
SCHOOL
HOME
HKD
HYP 3/5
INAT 6/9
IMPAIRMENT
20
ADHD versus HKD

• ANXIETY / DEPRESSION

IMP 1/4
SCHOOL
HOME
HKD
HYP 3/5
INAT 6/9
IMPAIRMENT
21
HKD AS A MODERATOR

• Using the same analysis as the original MTA
  report, HKD emerged as a significant moderator of
  treatment outcome on
• SNAP HYPERACTIVITY/IMPULSIVITY (P)
• SSRS TOTAL SOCIAL SKILLS SCORE (P)
• SSRS INTERNALISING SCORE (P)

22
SNAP Hyperactivity-Impulsivity (Parent)
23
SSRS Total Social Skills (Parent)
24
Treatment Decision Making - HKD
Hyperkinetic Disorder
STIMULANTS
25
Treatment decisions

• Severe, pervasive, disabling?
• Problems at home?
• Problems at school?
• Persistent after treatment?
• Comorbid problems?

Home CBT
?
Liaison
self-instruction
Medication
26
Medical treatment in comorbidity
As in simple ADHD more AEs enhanced monitoring
needed

• Mental retardation
• Autism spectrum
• Coordination problems
• Conduct disorders
• Anxiety
• Tourette
• Bipolar disorder
• Epilepsy
• Attachment disorder

Treat as usual
Predicts poor response but not contraindicated
Nonstimulants sometimes needed
Caution in BP I or II stimulants useful in
PBD
Stimulants are not contraindicated and may be
useful
Differentiate pattern cause does
not determine response
27
Predict by genetics?

• CYP2D6 polymorphisms for atomoxetine esterase
  for methylphenidate
• DRD4.7 findings contradictory
• DAT 10/10 may predict nonresponsiveness
• glutamate receptor 7 gene (GRM7) norepinephine
  transporter suggested in genome scan

Mick et al (2008) Am J Med Genet B
Neuropsychiatr Genet. 2008 51412-8.
Eg Keun-Ah Cheon, Young-Hoon Ryu, Jae-Won Kim
and Dae-YeonCho(2005)Eur Neuropsychopharm
15,,95-101
28
Stimulants raise dopamine levels
29
Atomoxetine raises frontal dopamine levels
By inhibiting the noradrenaline transporter
30
Adherence to Medication Regime (from baseline to
3 years in MTA trial)
31
Adherence and attitudes

• Tom is 15. Professional parents. White British.
• Mixed neuropsychiatric presentation
• Presented at age 10 with history of impulsive
  overactivity throughout his life asked to leave
  nursery multiple suspensions from primary school
  and three changes of school (all mainstream) due
  to mothers perception of school failing him
• Reading age then was 7 WISC IQ 106 noncompliant
  with tasks seen as difficult
• Increasingly unpopular steals to give to other
  kids
• Violent to his younger sister, not otherwise
• Treated with Concerta (in spite of tics
  appearing) good response, maintained in
  mainstream with facilitator, friendless.

32
Problems now

• Age 14 increasing cannabis use agreed to
  continue Concerta (54 mg daily) none the less
  off medication at weekends and holidays
  discussions in motivational interviewing format.
• Age 15 behaviour at school deteriorated. Concerta
  increased clonidine added not helpful admitted
  not taking medicines
• Wont accept a self-monitored trial dunno and
  dont like it on his objections.

33
Patients taking stimulants (General
Practice Research Database)
34
Common reasons for nonadherence

• Forget
• Stigma
• Not real self
• Losing funny side
• Adverse effects
• Physical sex tension feared brain damage
• Incompatible with misused substances
• Inconvenience
• Dont need it
• Up to me
• No point

35
Attitudes of young people to stimulants

• Harpur (2006, PhD thesis Southampton)
• Predominantly positive
• Adherence is complex individually chosen
  regimes, often by parents (Singh, 2006, Am J Med
  Ethics authenticity) adherence to what?
• Ferrin (2007, MSc, London)
• Questionnaire from Childrens Health Beliefs
  model locus of control, self-esteem, general
  beliefs on medicine, knowledge
  perceived threat and benefit
  doctor-patient relationship

36
Outcome and adherence

• Simpson et al BMJ 2006 333 15
• Metaanalysis good adherence in about 50
  predicts good outcome, even for placebo.
  (healthy adherer)
• Charach et al J Amer Acad CAP 43 559
• Adherence to stimulants over 5 years predicts
  good outcome, is predicted by youth, severity of
  ADHD, no ODD

37
Attitudes of young people to stimulants

• Project commissioned from LSE (Singh)
• Qualitative interviewing
• Attitudes predominantly positive
• Negative aspects acknowledged
• Inconvenient
• Stigmatising for some
• Sleep/appetite problems
• Better for some activities, worse for others

38
Conclusions

• Prediction of stimulant response depends upon the
  outcome desired.
• Greatest change happens in children with a
  pattern of inattentiveness, pervasive
  hyperkinesis and low anxiety
• Future possibilities for pharmacogenomics and
  fMRI
• Good compliance goes with good response and good
  communication