QEEG and MEDICATION

OUTLINE OF PRESENTATION

Jack Johnstone, Ph.D. [email protected]

Q-Metrx, Inc. July 20, 2007

I. Introduction

A. Disclosure

B. Scope of Presentation

II. Background

A. First, Do No Harm

B. Use of Antipsychotic Medications in Children

III. Powerful and Practical Applications of QEEG technology

A. Optimizing Clinical Outcome by Individualized Selection of Medications

B. Reduction of Side Effects

IV. Drug Effects on EEG: Principles

A. Quantity of medication in the patient influences the EEG. This quantity varies according to dose, volume of distribution, and rate of metabolism.

B. Combinations of medications (polypharmacy) may produce unpredictable effects

V. Pharmaco-EEG

A. Long history of study of effects of medications on EEG, as early as Hans Berger’s studies of effects of cocaine of beta frequency activity (Replicated in Herning et al. 1985.) see Herrmann, Itil, Fink, Saletu, Wauquier, others

B. Effects have now been studied in detail for many commonly used medications:

SEE: Saletu, B., Anderer, P., and G.M. Saletu-Zyhlarz, “EEG Topography and Tomography (LORETA) in the Classification and Evaluation of the Pharmacodynamics of Psychotropic Drugs” Clinical EEG and Neuroscience, April 2006, 66-80.

Chlorpromazine Haldol Imipramine Citalopram Clobazam Lorazepam d-Amphetamine Methylphenidate Pyritinol

C. General Effects of Medication on EEG

1. Psychostimulants (Ritalin, Dexedrine, Adderall, Concerta): Decreases slow activity (delta and theta frequencies) and increases fast activity (beta frequencies)

2. Benzodiazepines (Xanax, Valium, Ativan): Increases 14-25 Hz. Have anticonvulsant properties

3. Barbiturates (Phenobarbital): Increases delta activity and increases18-35 Hz beta spindles High dosage produces “burst-suppression”

4. Tricyclic Antidepressants: (Elavil, Tofanil, Norpramin, Sinequan, Pamelor) Increases both slow and fast activity and decreases alpha frequency activity (sedating)

5. SSRIs: (Prozac, Zoloft, Luvox, Paxil, Lexipro, Celexa) Produces less delta, decreases alpha, and increase beta (less sedating)

6. OTHER: NRIs (Wellbutrin- lowers seizure threshold); SNRIs (Cymbalta, Effexor); MAO Inhibitors (Nardil, Parnate), Increases 14-20 Hz beta (non-sedating)

7. Antipsychotics: 1) Tranquilizers (Thorazine, Haldol) Decreases fast activity, increases slow activity; 2) Atypical (Zyprexa, Risperdal, Seroquel, Clozaril) Increases delta, theta, increases beta.

8. Mood Stabilizers/Anticyclics: (Lithium, Tegretol) Increased theta frequency activity. Overdose produces marked slowing and triphasic discharges.

D. “Recreational” Drugs

1. Marijuana / Hashish / THC: Increased frontal alpha, increased interhemispheric alpha coherence.

2. Cocaine: Increased beta, EEG desynchronization, potentiates seizure activity

3. LSD: Decreased amplitude, suppression of slow activity, increases dominant frequency.

VI. Event-Related Potentials (ERPs)

A. Generally involves “signal averaging” of brief segments of EEG following sensory stimulation.

1. Visual, auditory, and somatrosensory stimuli are used to test afferent conduction through the central nervous system

2. “Cognitive” components can be identified that reflect psychological processes such as anticipation of stimuli or target detection of specific types of stimuli 3. The “P300” ERP is named because a POSITIVE potential following target stimulus can be identified peaking at 300 milliseconds post target stimulus.

4. The timing (latency) of the P300 appears to shift with age and is significantly delayed in dementing illness.

5. The latency of the P300 (and EEG spectra) can be normalized with the long-term use of anticholinesterase inhibitor such as Aricept (donezepil)

VII. The STAR*D Trial: “Sequenced Treatment Alternatives to Relieve Depression”

A. Largest and longest study ever done to evaluate treatment of depression

1. Over 4,000 patients enrolled in a “naturalistic” design

2. Four “levels” of treatment were studied (See attached START*D Treatment Flowchart)

B. About half of the participants became symptom-free after the first two treatment levels. After that, rates at which participants showed remission of symptoms slowed.

C. Over the course of ALL FOUR LEVELS, about 70 percent of those who did not withdraw from the study became symptom-free.

VIII. Using qEEG to PREDICT medication response

A. Suffin and Emory (1995) studies two groups of patients, one with affective disorder without attentional disturbance and one with attentional disorder without affective disturbance.

1. EEG Markers for response to stimulant medication were seen in each of the two patient groups (elevated frontal theta power)

2. EEG Markers for response to antidepressant medication were seen in each of the two patient groups (elevated frontal alpha)

B. R.J. Chabot et al., 1996, 1999 refined this work, examining effects of methylphenidate and Dexedrine.

1. Excess theta activity predicted response to Methylphenidate

2. Excess slow frequency alpha activity predicted response to Dexedrine.

C. Leuchter and Cook at UCLA developed a metric of the amount of relative power in the EEG in a given frequency range as a percentage of absolute power in the same frequency range termed “CORDANCE”.

1. Cordance in the theta frequency range showed changes with antidepressant medication. Changes were also seen placebo.

2. “This study demonstrates that although the symptomatic improvement resulting from placebo and medication treatment may be similar, the two treatments are not physiologically equivalent.”

IX. The BRITE-MD clinical trial (“STAR*D with qEEG”)

A. Following four independent pilot studies at UCLA, Cedars- Sinai Medical Center, and Mass General, Aspect Medical Systems initiated a large scale clinical trial at 10 sites.

B. A total of 375 patients with major depression were studies over a 13 week period. (See attached BRITE-MD protocol for details of study design)

C. A marker (frontal EEG) measured at one week following initiation of antidepressant medication (Lexipro) predicted clinical outcome at 7 weeks. Clinician prediction was at chance level.

D. If the change was NOT seen at one week it did not help to continue on Lexipro or augment Lexipro with Wellbutrin. Clinical outcome was improved by SWITCHING medication to Wellbutrin alone and discontinuing Lexipro (see attached poster).

X. EEG Phenotypes

A. Johnstone, et al. (2005) suggest that various EEG profiles might be considered as “intermediate phenotypes” falling between genetics and behavior.

B. EEG phenotypes may be considered as vulnerability factors and guide intervention for medications and neurofeedback.

C. Reprint available on request (contact [email protected])

XI. SUMMARY AND CONCLUSIONS

A. Numerous studies document the effects of various medications on EEG

B. Recent studies indicate that qEEG predicts medication response better than behavioral symptoms.

C. New technologies offer the promise of increased sensitivity and specificity.

Treatment Choices Throughout STAR*D

LEVEL1 – All participants were treated with citalopram (Celexa)

Those who went into remission (e.g., they became well) Went into follow-up

Those who did not get well, went on to Level 2

LEVEL 2 – Switching treatments or adding to citalopram (Celexa)

Those who chose to switch treatments were randomized to: • sertraline (Zoloft), Those who became well went into follow-up • bupropion-SR (Wellbutrin), • venlafaxine-ZR (Effexor), or • cognitive behavioral therapy (CBT)

Those who chose to add treatment were randomized to: Those who became well went into follow-up • bupropion-SR (Wellbutrin), • buspirone (BuSpar), or • cognitive behavioral therapy (CBT)

Those who did not get well went on to Level 3

LEVEL 3 – Switching treatments or adding to existing medication

Those who chose to switch treatments were randomized Those who became well went into follow-up to: • mirtazapine (Remeron) or • nortriptyline (Aventyl or Pamelor)

Those who chose to add treatment were randomized to: Those who became well • lithium or went into follow-up • triiodthyronine (T3)

Those who did not get well went on to Level 4 LEVEL 4 – Switching treatments

Participants were taken off all other medications and randomized to: • tranylcypromine, an MAOI (Parnate) or • venlafaxine XR (Effexor XR) + mirtazapine (Remeron)

Can EEG-guided Antidepressant Selection Improve Response Rates? Insights from the BRITE-MD Trial Andrew F. Leuchter, M.D.1, Ian A. Cook, M.D.1, William S. Gilmer, M.D.2, Scott D. Greenwald, Ph.D.3, Robert H. Howland, M.D.4, Madhukar H. Trivedi, M.D.5 1UCLA Laboratory of Brain, Behavior, and Pharmacology, Semel Institute, Los Angeles, CA, 2Psychiatry, Northwestern University, Chicago, IL,3Neuroscience, Aspect Medical Systems, Norwood, MA, 4Western Psychiatric Institute and Clinic, University of Pittsburgh, Pittsburgh, PA, 5Psychiatry, University of Texas Southwestern Medical School, Dallas, TX

ABSTRACT METHODS (Continued) RESULTS (Continued) Objective: The BRITE-MD study (www.BRITE-MD.org) was designed to assess ESC Group (n=40): 90% the accuracy of a frontal quantitative electroencephalographic (fqEEG) biomarker Continue escitalopram (10mg/day) 73% in predicting response to escitalopram (ESC) treatment. This analysis compares the 70% 68% 65% response rate between subjects who received treatment consistent with the 57% 52% 51% biomarker prediction vs. other subjects. 50% Method: 111 subjects (age: 42 ± 14 ; 59% female) meeting DSM-IV criteria for BUP Group (n=35): 41% 36% 35% MDD began treatment with escitalopram (ESC; 10 mg/day) and were randomly Switch to bupropion XL All subjects 30% assigned after 1 week to either: 1) continue ESC (10 mg/day; n=40); 2) augment (300 mg/day) (n=111) ent t nt with bupropion XL (AUG; 300 mg/day; n=36); or, 3) switch to bupropion XL jects e

start with Rates 7-Week Response b s onsis t t C en (BUP; 300 mg/day; n=35) for 7 weeks of treatment. At each visit severity of AUG Group (n=36): sistent nt escitalopram All Su n ubjec ist ts t ATR S s ste c depression was assessed with the Hamilton Depression Rating Scale (HAM-D-17) ll si e (10 mg/day) Continue escitalopram (10 mg/day) ATR Inconsist A bj ATR Co Incon on Su R ll and 4-channel fqEEG was recorded. Clinical response was defined as a reduction and add bupropion XL (300 mg/day) T A A ATR C in HAM-D at week 7 ≥ 50% from baseline. A previously developed index to ESC Group (n=40) ESC & BUP Groups (n=75) ESC, BUP & AUG Escitalopram GroupsATR Iinconsisten (n=111) predict probability of clinical response (0 to 100) using baseline and week 1 EEGs Fixed-Dose Treatment Phase (73% vs. 36%, p=0.025) (68% vs. 35%, p< 0.001) Challenge (Antidepressant Treatment Response (ATR rev 0.4)) was evaluated. Treatment (65% vs. 41%, p=0.008) consistent with ATR included subjects continued on ESC when ATR>=THRESHOLD or switched to alternate treatment when ATR< Figure 1. Potential Clinical Impact: Retrospective analysis of response Week 1 Week 7 THRESHOLD. All other subjects received treatment inconsistent with ATR. Baseline rates by consistency of treatment with ATR prediction demonstrated Randomize Endpoint higher response rates in subjects whose treatment was consistent with Discussion: For subjects remaining on the initial treatment (ESC), the response EEG and Clinician Prediction ATR prediction. rate was higher with ATR-consistent treatment vs. ATR-inconsistent treatment of Response (73% vs. 36%, p=0.025). For all subjects, the response rate was higher with ATR- consistent treatment than with ATR-inconsistent treatment (65% vs. 41%, CONCLUSIONS p=0.008), and higher than the pooled response rate (65% vs. 51%, p=0.043). ‹ ATR, an index (0 to 100) of EEG features from ‹ Antidepressant Treatment Response (ATR) Index at Conclusions: Using a fqEEG biomarker at week 1 of ESC treatment may help baseline and week1 recordings, was prospectively week 1 of ESC treatment may help guide antidepressant guide antidepressant selection. Subjects whose ATR predicts response do better evaluate to estimate the probability of clinical response when continued on ESC, while subjects whose ATR predicts non-response may selection benefit from alternate regimens. ‹ Subjects treated consistently with ATR prediction Funding: This research was supported by Aspect Medical Systems, Inc. were those who continued on ESC when ATR ≥ ¾ Subjects whose ATR predicts response do better when THRESHOLD and those switched to alternate treatment continued on ESC, while subjects whose ATR predicts INTRODUCTION when ATR < THRESHOLD. All others subjects received non-response may benefit from alternate regimens treatment inconsistent with ATR. ‹ Prior work demonstrated that a simple-to-use frontal ¾ When ATR predicts non-response, switching to BUP, quantitative EEG (fqEEG) biomarker (ATR) predicted rather than augmenting with BUP, may be the preferred RESULTS response to antidepressant treatment in MDD [1][2] alternate regimen ‹ 111 subjects completed 7 weeks of treatment (age: ‹ This work compares response rates between subjects ¾ Clinical implication: Early identification of positive or treated consistent with biomarker prediction vs. other 42 ± 14; 59% female) negative EEG response to treatment may aid in decisions subjects in an interim analysis of the BRITE-MD trial ‹ For all subjects, the response rate was higher with regarding medication adjustments, potentially leading to (www.BRITE-MD.org) ATR-consistent treatment than with ATR-inconsistent treatment (65% vs. 41%, p=0.008) and higher than improved outcomes of antidepressant therapy METHODS the pooled response rate (65% vs 51%, p=0.043) ‹ MDD subjects (DSM-IV criteria; baseline IDS-SR >12) ‹ For subjects remaining on the initial treatment REFERENCES began treatment with escitalopram and were randomly (ESC), the response rate was higher for ATR-consistent assigned after 1 week to either: 1) continuation of treatment vs. ATR-inconsistent treatment (73% vs. [1] Iosifescu D, Greenwald S, Smith C, Devlin P, Alpert J, Hamill S, Fava escitalopram (ESC;10mg/day), 2) augment with 36%, p=0.025) M. Frontal EEG at 1 Week Predicts Clinical Response to SSRI Treatment in Major Depressive Disorder . Presented at the 2006 bupropion XL (AUG; 300 mg/day), or 3) switch to ‹ For subjects receiving alternate treatment and bupropion XL (BUP;300mg/day) for a total of 7 weeks Annual Meeting of the American Psychiatric Association, Toronto, CA whose ATR predicted non-response, those receiving (#231). ‹ At each study visit, 4-channel fqEEG was recorded BUP trended towards a higher response rate than [2] Poland R, Greenwald S, Smith C, Kustak C, Schulz J, Rowe S, and HAM-D-17 was assessed. Clinical response was those with augmentation (AUG) (58% vs. 50%, defined as a reduction in HAM-D-17 ≥ 50% at week 7 Gertsik L. Frontal EEG at 1 Week Predicts Response to Treatment p=n.s.) with Citalopram in MDD. Presented at the 2006 Annual Meeting of the from baseline American Psychiatric Association, Toronto, CA (#269). 2007 APA #284