WHAT CAN WE LEARN FROM STAR*D (Sequenced Treatment Alternatives to Relieve Depression) Ira Lesser, M.D. Chair, Department of Psychiatry Harbor-UCLA Medical Center Professor, Department of Psychiatry and Biobehavioral Sciences Geffen School of Medicine at UCLA

DISCLOSURES Grant support National Institute for Mental Health Bristol-Myers Squibb Forest Pharmaceuticals Aspect Medical Systems

Disclosures None of my slides and/or handouts contain any advertising, trade names or product- group messages. Any treatment recommendations I make will be based on clinical evidence or guidelines. Ira Lesser, M.D. Harbor-UCLA Medical Center

Focus of Presentation Discuss consequences of untreated/partially treated depression Discuss treatment resistant depression Discuss research approaches to study this, e.g. efficacy vs effectiveness trials Discuss the STAR*D trial and methodology Discuss the STAR*D results and implications for clinical practice

Health Burden Of Depression MDD is common and recurrent and can be disabling Lifetime prevalence ~ 10-15% Women are affected more than men Over 2/3 of people have recurrences Depressed adults have twice the annual health care costs as non-depressed World-wide is the 4 th most disabling medical condition, climbing to 2 nd by 2020

Definitions of Response and Remission % Reduction in Score Remission>75% Response50% - 74% Partial Response25% - 49% Nonresponse<25%

Nonremission is Common 35–45% remission 10–20% response with residual symptoms 15% partial response 25% nonresponse 7–15% intolerant Depression in Primary Care, Vol. 2. Treatment of Major Depression. Rockville, MD: US Dept. of Health and Human Services, AHCPR Publication No , 1993.

Staging Treatment Resistance Stage I Inadequate response to 1 monotherapy Stage II Inadequate response to 2 adequate monotherapy trials (different classes) Stage III Stage II resistance plus inadequate response to 1 augmentation trial Stage IV Stage III resistance plus inadequate response to a second augmentation trial Stage V Stage IV resistance plus inadequate response to bilateral ECT Adapted from: Thase ME, Rush AJ. J Clin Psychiatry. 1998;59(suppl 5):5 Souery D et al. Eur Neuropsychopharmacol. 1999;9:83

Pharmacological Options After Failure of First Antidepressant Optimize dose and address adherence Change to another antidepressant Same class Different class Add a second antidepressant Add a non-antidepressant Lithium or other mood stabilizer Thyroid hormone Psychostimulant Atypical antipsychotic

Efficacy vs Effectiveness Randomized Clinical Trials EFFICACY Aims for pure populations Assesess safety and efficacy Co-morbid conditions excluded Rates for MDD response are about 50%, 20-30% remission EFFECTIVENESS Looks for “real world” subjects Assesess effectivenenss Co-morbid conditions are OK Rates for MDD response are low

The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) Trial Rationale and Design ( A. John Rush, M.D. University of Texas Southwestern Medical Center Dallas, Texas

Importance of STAR*D The largest clinical trial of depression ever Conducted in primary care as well as psychiatric settings Few exclusion criteria, making it “real world” Included large numbers of minority patients Included cognitive therapy Combined randomization and patient choice The outcome was “remission” rather than “response”

Why Remission Was the End Point Remitters have less disability, better role function, life satisfaction, and less recurrences

Consequences of Nonremission Poor function (e.g., work, family) Poor prognosis (e.g., increased recurrence) Psychiatric or general medical complications (e.g., substance abuse) Health service utilization Death from:  Medical comorbidities  Suicide Treatment resistance

*Relapse defined as onset of new major depressive episode. Adapted from Judd LL et al. J Affect Disord. 1998;50: Median Weeks to Relapse* Following Response Improved Without Remission (n = 82 ) Remission (n = 155) Weeks

STAR*D Overview - I Duration: 7 years (October September 2006) Funding: National Institute of Mental Health National Coordinating Center, UT Southwestern Medical Center, Dallas Data Coordinating Center, Pittsburgh

STAR*D Overview - II 14 Regional Centers 41 Clinical Sites 18 Primary Care Settings 23 Psychiatric Care Settings

Overall Aim of STAR*D Define preferred treatments for patients who failed one SSRI All subjects begin on citalopram Doses are maximized Remitters enter follow-up If no remission, go to level 2 and subsequent levels

Participants Major depressive disorder Nonpsychotic Representative primary and specialty care practices (nonacademic) Self-declared patients

Inclusion Criteria Clinician deems antidepressant medication indicated years of age. Baseline HRSD 17  14. Most concurrent Axis I, II, III disorders allowed.

Multiple Research Outcomes Symptoms Function Side effect burden Patient and clinician satisfaction Utilization and costs of health care services

Clinical Procedures Open treatment with randomization Symptoms/side effects measured at each clinical visit Clinicians guided by algorithms/ supervision Dose adjustments”mandatory” to achieve remission (QIDS-C 16 ) Education for all patients

Level 1 Findings

Demographic Baseline Features (N=2876) Age (yrs.)40.8 (13.0) % Female63.7 Race % White75.8 % African-American17.6 % Others6.6 % Latino13.0 Age Groups years years years25.8 Education (yrs.)13.4 (3.2) Trivedi et al., Am J Psychiatry, 163(1):28-40, 2006

Education % < High School12.6 High School < College62.2 > College25.2 Insurance Private 51.1 Public 14.2 None 34.7 % Primary Care37.9 Mean (SD) Household Income ($ mo.)2358 (3030) Demographic Baseline Features (N=2876) Trivedi et al., Am J Psychiatry, 163(1):28-40, 2006

Marital Status % Never Married28.7 Married41.7 Divorced26.5 Widowed3.1 Employment Status Employed56.2 Unemployed38.2 Retired5.6 Social Baseline Features (N=2876) Trivedi et al., Am J Psychiatry, 163(1):28-40, 2006

% Recurrent Depression75.7 Onset age < 18 yrs.37.8 Positive Family History of Depression55.5 History of Attempted Suicide17.9 Current MDE > 24 months25.3 Anxious Depression53.2 Clinical Baseline Features (N=2876) Trivedi et al., Am J Psychiatry, 163(1):28-40, 2006

Mean (SD) HRSD 17 (ROA) 21.8 (5.2) QIDS-SR (4.0) Age at First Onset (yrs.)25.3 (14.4) # of MDEs6.0 (11.4) Length of Current MDE (mos.)24.6 (51.7) Length of Illness (yrs.)15.5 (13.2) Clinical Baseline Features (N=2876) Trivedi et al., Am J Psychiatry, 163(1):28-40, 2006

Majority Had Concurrent Axis I Disorders at Baseline (N=2876) # Concurrent Axis I Disorders % Trivedi et al., Am J Psychiatry, 163(1):28-40, 2006

% Generalized Anxiety Dis.23.6 Obsessive Compulsive Dis.14.3 Panic Disorder13.1 Social Phobia 31.3 Posttraumatic Stress Dis.20.6 Agoraphobia 11.8 Alcohol Abuse/Dependence12.1 Drug Abuse/Dependence 7.4 Somatoform Dis.2.4 Hypochondriasis4.4 Bulimia13.0 Trivedi et al., Am J Psychiatry, 163(1):28-40, 2006 Concurrent Baseline Axis I Disorders a (N=2876) a Defined by PDSQ.

Treatment and Symptom Outcomes (n=2876) Response (without remission) 50% decrease in baseline QIDS-SR 16 Remission HAMD 17 < 7 QIDS-SR 16 < 5 Citalopram Dose (at level exit): 41.8 mg (S.D. 16.8) Duration: 10.0 weeks (S.D. 4.2) Trivedi et al., Am J Psychiatry 2006;163:28-40

Treatment Outcome: Level 1 (N=2876) HAMD-17=17-item Hamilton Rating Scale for Depression QIDS-SR-16=16-item Quick Inventory of Depressive Symptomatology – Self-Report Trivedi et al., Am J Psychiatry 2006;163:28-40

Similar Outcomes in Primary and Psychiatric Care Settings (N=2876) % Trivedi et al., Am J Psychiatry 2006;163:28-40

Remission vs. Non-remission Remitters stayed in treatment longer (12 vs. 9.3 weeks) Remitters stayed on final dose longer (6.6 vs. 4.4 weeks) Remitters had less side effect burden

Of Ultimate Remitters, 1/2 Remitted by Week 6 n=2,876 Remission= QIDS-SR 16 < 5 Trivedi et al., Am J Psychiatry, 163(1):28-40, %

Remission vs. Non-Remission: Significant Baseline Differences Female gender Being employed Caucasian (vs. African- American) Being married or co- habitating Being more educated Having private insurance Having less medical problems Having less psychiatric co-morbidities Lower baseline severity Better baseline physical and mental function Greater life satisfaction Shorter current episode

Summary: Level 1 Over one-third of patients have been depressed for >2 years and 2/3 have concurrent GMCs About 1/3 will remit Response occurs in ~1/2 AFTER 6 weeks Measurement Based Care is feasible and works Studies of remission require longer study periods than 8 weeks When protocol-based care is given, results are equivalent in primary and specialty care

Level 2 Findings

Randomize SERBUP-SRVEN-XRCT CIT + BUP-SR CIT + BUS CIT + CT Level 2 Augmentation Options Switch Options

Acceptability of Treatment Options 41% chose only to have medication switch 30% chose only to have medication augmentation 11% chose to have any augmentation (meds or CBT) 7% chose to have any switch (meds or CBT) 1.4% chose to have any available option

Level 2 Medication Switch

Treatment Outcomes: Level 2 Switch (% remission) Rush et al., N Engl J Med 2006;354(12): (N-239)(N = 238)(N = 250)

Level 2 Medication Augmentation

Treatment Outcomes: Level 2 Augmentation (% remission) (N = 279) (N = 286) Trivedi et al., N Engl J Med 2006;354(12):

Summary: Level 2 (switch) Mean doses/day: bupropion=282.7mg; sertraline=135mg; venlafaxine=193mg ~ 25% achieve remission after switching to another antidepressant No significant differences among various antidepressants in achieving remission: changing class of medication did not make a difference Intolerance to citalopram did not predict intolerance to sertraline Perhaps venlafaxine dose too low

Summary: Level 2 (augment) Mean doses/day: bupropion=267mg; buspirone=41mg ~ 30% achieve remission augmenting citalopram, with no absolute difference between treatments Citalopram plus bupropion led to greater degree of improvement across subjects Bupropion was better tolerated

Level 2 Cognitive Therapy Findings

Cognitive Therapy Arm Certification process for CT therapists Session twice weekly for weeks 1-4, then once weekly for remaining 8 weeks (16 visits) If there was response without remission, could offer additional 8 sessions 469/1439 (26%) of those eligible chose CT as a possible option 65 switched to CT; 36 augmented with CT

Treatment Outcomes (% Remission) (L-2 CT vs. Med Switch) (N = 36) (N = 86) Thase et al., in preparation

Treatment Outcomes (% Remission) (L-2 CT vs. Med Augment) Thase et al., in preparation (N = 65) (N = 117)

CT Conclusions CT is both an acceptable switch and augmentation option in the second step Benefit of CT as augmentation was slower (up to 3 weeks) compared to augmenting with medication Whether CT responders/remitters fare better in follow-up is to be analyzed CT was not as popular as expected, thus limiting statistical power

Level 3 MRTNTP L-2 Tx + Li L-2 Tx + THY SwitchAugmentation Randomize

Level 3 Treatment Features Medication switch Mirtazapine mean final dose was 42.1 mg Nortriptyline mean final dose was 96.8 mg Medication augmentation Lithium mean final dose was 860 mg T 3 mean final dose was 45 μg

Treatment Outcomes: Level 3 Switch (% Remission) (N = 114) (N = 121) Fava et al., Am J Psychiatry, in press

Treatment Outcomes: Level 3 Augmentation (% remission) (N = 69) (N = 73) Nierenberg et al., Am J Psychiatry, submitted

Summary: Level 3 (switch) Remission rates (<20%) and side effect profiles with nortriptyline and mirtazapine were not different Low remission rates raises questions about sequential use of single agents after two failed trials, even with different pharmacologic action

Summary: Level 3 (augment) Remission rates with switching was modest, at best Remission rates augmenting with Li or Thyroid hormone were not different Questions as to whether lithium was adequately dosed T 3 augmentation was better tolerated

Level 4 TCP VEN-XR + MRT Randomize Switch

Treatment Outcomes: Level 4 (% Remission) (N = 58) (N = 51) McGrath et al., Am J Psychiatry, submitted

Summary: Level 4 Mean doses/day:Tranylcypromine=37mg; Venlafaxine=210mg; Mirtazapine=36mg Dosing was lower than might have been expected Tranylcypromine required a washout and had higher early dropouts Overall remission rates were low, with no difference between treatments VEN-XR/MIRT had some advantages (tolerability) over TCP

Remission Rates by Levels a a By QIDS-SR 16 <5 at level exit Level 1 (2876) 32.9 Level 2 (1439) Switch (789) Switch (789) Augment (650) Augment (650) Level 3 (377) Switch (235) Switch (235) Augment (142) Augment (142) Level 4 (109) 14.7

Follow-up Findings

Follow up Follow up was naturalistic; visits recommended every two months Patient were told to continue what was effective before, but any medication dose change or psychotherapy was allowed Response and remission defined by QIDS-SR score as in acute phase; relapse defined by QIDS-SR > 11 (~ equal to HAM-D of 14)

Level 1 Follow-Up Relapse = QIDS-IVR 16 > 11

Level 2 Follow-Up Relapse = QIDS-IVR 16 > 11

Level 3 Follow-Up Relapse = QIDS-IVR 16 > 11

Level 4 Follow-Up Relapse = QIDS-IVR 16 > 11

p<.0001 Relapse = QIDS-IVR 16 > Step2 Steps3 Steps4 Steps (N=1085) (N=383) (N=35) (N=15) Relapse in Follow-up for Participants Remitting with Different Numbers of Acute Treatment Steps

Relapse in Follow-up for Participants Not remitting to Different Numbers of Acute Treatment Steps p<.0001 Relapse = QIDS-IVR 16 > Step2 Steps3 Steps4 Steps (N=388) (N=237) (N=66) (N=34)

Summary: Follow Up Failure to reach remission during acute treatment portends a worse prognosis Remitters at entry to F/U had better prognosis than those who improved, but did not remit Relapse rates were higher for those who entered F/U after more acute treatment steps For those who relapse, mean time to relapse was shorter for those who required >2 steps

Conclusions Cumulative remission rate is over 50% with first 2 steps Both within-class and out-of-class switches are equally effective (level 2) The duration of acute treatment needed to see response was 8-10 weeks With careful symptom and medication review results are similar in primary and specialty care Patient preference plays a big role in strategy selection

Conclusions Pharmacologic distinctions do not lead to large clinical differences Cognitive therapy is viable switch or augmenting option, though many chose not to partake in it It is important not to change course too quickly nor to give up on multiple sequential strategies Psychosocially disadvantaged patients did more poorly and may need innovative approaches

Future Directions Further analyze the cognitive therapy data Report multiple sub-analyses We need newer approaches to the treatment refractory patient Develop new medications and treatment protocols to address non-remission, e.g. should patients with chronic depression be started on two medications simultaneously? What about combination medication and psychotherapy approaches?

The STAR*D Study Investigators National Coordinating Center A. John Rush, MD Madhukar H. Trivedi, MD Diane Warden, PhD, MBA Melanie M. Biggs, PhD Kathy Shores-Wilson, PhD Diane Stegman, RNC Michael Kashner, PhD, JD Data Coordinating Center Stephen R. Wisniewski, PhD G.K. Balasubramani, PhD James F. Luther, MA Heather Eng, BA. University of Alabama Lori Davis, MD University of California, Los Angeles Andrew Leuchter, MD Ira Lesser, MD Ian Cook, MD Daniel Castro, MD University of California, San Diego Sidney Zisook, MD Ari Albala, MD Timothy Dresselhous, MD Steven Shuchter, MD Terry Schwartz, MD Northwestern University Medical School, Chicago William T. McKinney, MD William S. Gilmer, MD

The STAR*D Study Investigators University of Kansas, Wichita and Clinical Research Institute Sheldon H. Preskorn, MD Ahsan Khan, MD Massachusetts General Hospital, Boston Jonathan Alpert, MD Maurizio Fava, MD Andrew A. Nierenberg, MD University of Michigan, Ann Arbor Elizabeth Young, MD Michael Klinkman, MD Sheila Marcus, MD New York State Psychiatric Institute and Columbia College of Physicians and Surgeons, New York Frederic M. Quitkin, MD Patrick J. McGrath, MD Jonathan W. Stewart, MD Harold Sackeim, PhD University of North Carolina, Chapel Hill Robert N. Golden, MD Bradley N. Gaynes, MD

The STAR*D Study Investigators Laureate Healthcare System, Tulsa Jeffrey Mitchell, MD William Yates, MD University of Pittsburgh Medical Center, Pittsburgh Michael E. Thase, MD Edward S. Friedman, MD Vanderbilt University Medical Center, Nashville Steven Hollon, PhD Richard Shelton, MD The University of Texas Southwestern Medical Center, Dallas Mustafa M. Husain, MD Michael Downing, MD Diane Stegman, RNC Laurie MacLeod, RN Virginia Commonwealth University, Richmond Susan G. Kornstein, MD Robert K. Schneider, MD

Pharmaceutical Industry Support STAR*D medications were provided gratis by Bristol-Myers Squibb Company, Forest Pharmaceuticals Inc., GlaxoSmithKline, King Pharmaceuticals, Organon Inc., Pfizer Inc., and Wyeth-Ayerst Laboratories.

References Rush AJ, Fava M, Wisniewski SR et al: Sequenced Treatment Alternatives to Treat Depression (STAR*D): rationale and design. Control Clin Trials 2004; 25: Trivedi MH, Rush AJ, Wisniewski SR et al: Evaluation of outcomes with citaloprram for depression using measurement-based care in STAR*D: Implications for clinical practice. Am J Psychiatry 2006; 163:1-13 Rush AJ, Trivedi M, Wisniewski SR et al: Bupropion-SR, sertraline, or Venlafaxine-XR after failure of SSRIs for depression. N Engl J Med 2006;354: Trivedi MH, Rush AJ, Wisniewski SR et al: Medication augmentation after the failure of SSRIs for depression. N Engl J Med 2006;354:

References Fava M, Rush AJ, Wisniewski SR et al: A comparison of mirtazapine and nortriptyline following two consecutive failed medication treatments for depressed outpatients: a STAR*D report. Am J Psychiatry 2006;163: Insel T: Beyond efficacy: the STAR*D Trial. Am J Psychiatry 2006; 163:5-7 Nierenberg AA, Fava M, Trivedi MH et al: A comparison of lithium and T 3 augmentation following two failed medication treatments for depression: a STAR*D report. Am J Psychiatry 2006;163: McGrath PJ, Stewart JW, Fava M et al: Tranylcypromine versus venlafaxine plus mirtazapine following three failed antidepressant medication trials for depression: A STAR*D report. Am J Psychiatry 2006:163:

References Rush AJ, Trivedi MH Wisniewski SR et al: Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry 2006;163: Rubinow DR: Treatment strategies after SSRI failure—good news and bad news. N Engl J Med 2006;354: Valenstein M: Keeping our eyes on STAR*D. Am J Psychiatry 2006;163: Menza M: STAR*D: the results begin to roll in. Am J Psychiatry 2006; 163: Nelson CJ: The STAR*D study: a four-course meal that leaves us wanting more. Am J Psychiatry 2006;163: Rush AJ: STAR*D: What have we learned? Am J psychiatry 2007;164: Insel T: Beyond efficacy: the STAR*D Trial. Am J Psychiatry 2006; 163:5-7