2007 4306s1 10 FDA Egan

ZIMULTISANOFI-AVENTISNDA 21-888 : 

ZIMULTI SANOFI-AVENTIS NDA 21-888 Endocrinologic and Metabolic Drugs Advisory Committee Meeting Silver Spring, Maryland June 13, 2007 Amy G. Egan, M.D., M.P.H. Division of Metabolism and Endocrine Products Center for Drug Evaluation and Research

Outline: 

Outline Efficacy findings and conclusions for the indication of weight management Safety concerns Neurological adverse events Seizures Psychiatric adverse events Suicidality

FDA Guidance Criteria for Efficacy of Weight-Management Drugs: 

FDA Guidance Criteria for Efficacy of Weight-Management Drugs The drug’s effect is significantly greater than that of placebo with the mean drug-associated weight loss exceeding mean placebo weight loss by at least 5%. The proportion of subjects who reach and maintain a loss of at least 5% of their initial body weight is significantly greater in subjects on drug than in those on placebo.

Efficacy Conclusions: 

Efficacy Conclusions Rimonabant 20 mg daily along with a hypocaloric diet was shown to reduce body weight by approximately 5% relative to hypocaloric diet plus placebo Rimonabant-associated weight loss was accompanied by improvement in levels of triglycerides, HDL-C, and HbA1c in subjects with type 2 DM

Efficacy Conclusions (cont’d): 

Efficacy Conclusions (cont’d) Relative to placebo, rimonabant had no effect on levels of total cholesterol or LDL-C For unclear reasons, reductions in systolic and diastolic blood pressure were less than expected given the degree of weight loss

Caveats: 

Caveats High attrition Year 1 drop out rates: ~ 32% to 49% Year 2 drop out rates: ~ 23% to 58% No systematic follow-up of dropouts Last observation carried forward (LOCF) was used in this study as the 1° analytic method Generalizability Middle-aged Caucasian females Greater efficacy in Caucasian vs. African-American and younger vs. older Subjects with a history of significant depression excluded

Caveats: 

Caveats High attrition Year 1 drop out rates: ~ 32% to 49% Year 2 drop out rates: ~ 23% to 58% No systematic follow-up of dropouts Last observation carried forward (LOCF) was used in these studies as the 1° analytic method Generalizability Middle-aged Caucasian females Greater efficacy in Caucasian vs. African-American and younger vs. older Subjects with a history of significant depression excluded

Caveats: 

Caveats High attrition Year 1 drop out rates: ~ 32% to 49% Year 2 drop out rates: ~ 23% to 58% No systematic follow-up of dropouts Last observation carried forward (LOCF) was used in this study as the 1° analytic method Generalizability Middle-aged Caucasian females Greater efficacy in Caucasian vs. African-American and younger vs. older Subjects with a history of significant depression excluded

Rimonabant Database: 

Rimonabant Database As of 18 December 2006, the cumulative database for rimonabant consists of: 1308 healthy subjects from 37 completed Phase 1 studies 1230 patients from 5 completed Phase 2 studies 13,644 patients from 13 completed Phase 3 studies 6761 obese patients in 8 weight management/ diabetes studies 6883 patients in 5 smoking cessation studies)

Exposure to Rimonabant: 

Exposure to Rimonabant Subject exposure to rimonabant 20 mg indications as of December 18, 2006: 0 to < 6 months: 2256 subjects 6 to <12 months: 1619 subjects 12 to < 18 months: 592 subjects 18 to 24 months: 441 subjects

Selection of Events of Interest: 

Selection of Events of Interest Datasets: Adverse event dataset Associated psychiatric symptoms dataset Controlled substance dataset Studies with re-randomization: Same treatment during the entire study

Statistical Methodology: 

Statistical Methodology Meta-analysis stratified by study Included 14 randomized phase 2 and 3 trials Per study sample sizes ranged from 20 to 3,000 per group Study durations ranged from 4 to 104 weeks Primary treatment comparison: rimonabant 20 mg vs. placebo

Statistical Methodology: 

Statistical Methodology RR, OR, RD Safety outcomes with relatively rare events (seizure, suicidality): an exact meta-analysis and a fixed-effects meta-analysis were used Safety outcomes with more common events (neurological AEs and psychiatric AEs): Fixed and random effects meta-analyses were used Primary analysis used first randomization data only Sensitivity analyses were performed using second randomizations

Neurological Adverse Events: 

Neurological Adverse Events

Endocannabinoids and the Nervous System: 

Endocannabinoids and the Nervous System CB1 receptors abundant in the cerebellum, cortex, hippocampus, hypothalamus, and basal ganglia (involved in memory, motor function and reward behaviors) CB1 receptors also present on the peripheral nerves Neuroprotection of CNS and PNS

Neurological Adverse Events: 

Neurological Adverse Events Occurred in 27.4% of rimonabant 20 mg treated subjects vs. 24.4% of placebo treated subjects Responsible for 3.5% of discontinuations in the rimonabant 20 mg treated subjects vs. 1.3% of placebo treated subjects

Sensory Changes: 

Sensory Changes

Motor Impairment: 

Motor Impairment

Cognitive Disorders: 

Cognitive Disorders

Relative Risk for Neurological Adverse EventsRIO Studies: 

Relative Risk for Neurological Adverse Events RIO Studies

Relative Risk for Neurological Adverse Events Diabetes Studies: 

Relative Risk for Neurological Adverse Events Diabetes Studies

Multiple Sclerosis: 

Multiple Sclerosis Clinical trials data: Cannabinoids can reduce the spasms, spasticity, or tremor of MS Pre-clinical data: Mouse models of MS suggest that CB receptor activation may oppose the progression of MS Slows neurodegenerative process Reduces inflammation Promotes remyelination

Multiple Sclerosis: 

Multiple Sclerosis Numbers small to date RIO 2 confirmed MS cases on Rim 5 mg 1 un-confirmed MS case on placebo Smoking cessation 2 suspected cases on Rim 20 mg Post-marketing 1 case of optic neuritis; MRI suggestive of MS 1 case in subject with MS with exacerbation on Rim Lag time to diagnosis Initial vague symptomatology Theoretic possibility Biologically plausible

Seizures: 

Seizures

Endocannabinoids and Seizure Potential: 

Endocannabinoids and Seizure Potential Endocannabinoid system has been implicated in regulating seizure duration and frequency. Seizure activity elicits an increase in endocannabinoids which activate pre-synaptic CB1 receptors with subsequent regulation of neuronal hyperexcitability and seizure termination

Pre-Clinical Data Supporting Seizure Risk: 

Pre-Clinical Data Supporting Seizure Risk 6% of rats and mice and 20% of monkeys developed seizures while receiving long-term treatment with doses of rimonabant 0.5-2 times the 20 mg dose versus 1.5% of control mice and no control rats or monkeys Rimonabant accumulates 2-fold in the brain with multiple dosing

Exclusion Criteria Relevant to the Evaluation of Seizure: 

Exclusion Criteria Relevant to the Evaluation of Seizure Presence of any clinically significant neurological disease Presence of treated epilepsy Prolonged administration (> 1 week) of neuroleptics within 3 months prior to screening visit Subjects discontinued from the trials if treated with a neuroleptic

Cases of Seizure: 

Cases of Seizure

Relative Risk for All Reported Seizures: 

Relative Risk for All Reported Seizures

Seizures in Ongoing Studies: 

Seizures in Ongoing Studies 8 cases of seizure in ongoing studies 6 on rimonabant 20 mg 2 on placebo Ongoing studies all have a 1:1 randomization

Seizure Conclusions: 

Seizure Conclusions Biological plausibility linking rimonabant to seizure risk Pre-clinical data indicate proconvulsant effect of rimonabant Exclusion criteria screened out high-risk individuals from clinical trials Although the current clinical data perhaps suggest an increased seizure risk with rimonabant, only clinical experience will clarify this potential risk

 Psychiatric Adverse Events: 

Psychiatric Adverse Events

Endocannabinoids and Psychiatric Disorders : 

Endocannabinoids and Psychiatric Disorders Endocannabinoids act as modulators in pathological conditions Anxiety, phobia, depression, post-traumatic stress disorder CB1 receptors are abundant in the pre-frontal cortex Regulation of mood, aggression and/or impulsivity and decision making CSF levels of the endogenous cannabinoid anandamide correlate inversely with psychotic symptoms in schizophrenic patients

Exclusion Criteria Relevant to the Evaluation of Psychiatric Adverse Events: 

Exclusion Criteria Relevant to the Evaluation of Psychiatric Adverse Events Presence of any clinically significant psychiatric disease History of severe depression defined as: Depression necessitating hospitalization or; History of 2 or more recurrent episodes of depression or; History of suicide attempt or; Prolonged administration (more than 1 week) of anti-depressants within 3 months prior to screening Subjects discontinued from the trials if started on an anti-depressant 11,225 subjects were screened for RIO; 113 (1%) failed screening due to these exclusion criteria

Baseline History of Depressed Mood Disorders and Disturbances: 

Baseline History of Depressed Mood Disorders and Disturbances Ongoing at randomization: 1.4% on rimonabant 20 mg vs. 1.1% on placebo

Psychiatric Adverse Event Terms: 

Psychiatric Adverse Event Terms Primary System Organ Class (SOC): Psychiatric Disorders High Level Group Terms (HLGT) Preferred Terms (PT) Anxiety Disorders and Symptoms (anxiety, nervousness, stress, tension) Depressed Mood Disorders and Disturbances (depression, depressed mood, anhedonia, dysthymic disorder) Sleep Disorders and Disturbances (insomnia, parasomnia, somnolence) Mood Disorders and Disturbances (affect alteration, crying, mood altered)

Psychiatric Adverse Event Rates by High Level Group Term (RIO): 

Psychiatric Adverse Event Rates by High Level Group Term (RIO)

Relative Risk of Psychiatric AE (RIO): 

Relative Risk of Psychiatric AE (RIO)

Relative Risk of Psychiatric Adverse Event by Subgroups: 

Relative Risk of Psychiatric Adverse Event by Subgroups Age 65 years or older RR = 3.1 Less than 65 years RR =1.7 Geographic location US studies RR = 1.7 Non-US studies RR = 1.8

Relative Risk of Psychiatric Adverse Event – by Subgroups : 

Relative Risk of Psychiatric Adverse Event – by Subgroups Gender Males RR = 2.1 Females RR = 1.7 5% Weight-Loss Response Responders RR = 1.9 Non-responders RR=1.9

Psychiatric Adverse Event by Baseline History of Depressed Mood (RIO): 

Psychiatric Adverse Event by Baseline History of Depressed Mood (RIO) 1082/1235 (88%) of subjects who experienced a psychiatric adverse event did not have a baseline history of Depressed Mood Disorders and Disturbances

Treatment of Psychiatric Adverse Events: 

Treatment of Psychiatric Adverse Events

Psychiatric Adverse Events Conclusions: 

Psychiatric Adverse Events Conclusions Rimonabant 20 mg was associated with an ~ 2-fold increase in the risk of psychiatric adverse events Anxiety disorders and symptoms Depressed mood disorders and disturbances Sleep disorders and disturbances

Suicidality: 

Suicidality

Columbia Suicidality Categories : 

Columbia Suicidality Categories Category 1 Completed suicide 2 Suicide attempt 3 Preparatory acts toward imminent suicide behavior 4 Suicidal ideation 5 Self-injurious behavior, intent unknown 6 Not enough information (fatal) 9 Not enough information (non-fatal)

Suicidality Assessment: 

Suicidality Assessment 1201 patient narratives assessed by the Posner group 91 suicidality cases identified: Definitely (Columbia categories 1, 2, 3, or 4) Possibly (Columbia categories 5, 6, or 9) Majority of cases were assessed suicidal ideation (Columbia Category 4): Placebo: 14 Rim 5 mg: 10 Rim 20 mg: 40

Statistical methodology: 

Statistical methodology 13 studies used in the analysis RIO North America and EFC4796 re-randomized patients First randomization used in primary analysis EFC4796 used the Rim 5 mg group as the control group (no placebo group in first randomization) Sensitivity analyses performed for the 2nd randomization

Suicidality Cases in the Analyses: 

Suicidality Cases in the Analyses Total number of suicidality cases contributing to the analyses is 74

Odds Ratio of SuicidalityAll Studies : 

Odds Ratio of Suicidality All Studies

Odds Ratio of SuicidalityObesity Studies: 

Odds Ratio of Suicidality Obesity Studies

Completed Suicides: 

Completed Suicides Four completed suicides reported to date - 3 in entire rimonabant clinical trial database; 1 post-marketing All three in clinical trials have occurred in subjects taking rimonabant: RIO North America: 63 year-old male on Rim 5 mg STRADIVARIUS: 36 year-old male on Rim 20 mg CRESCENDO: 77 year-old male on Rim 20 mg Post-marketing: 33 year-old male on Rim 20 mg

Rimonabant and SuicidalityWhat is the Nature of the Association? : 

Rimonabant and Suicidality What is the Nature of the Association? Bias Ascertainment Confounding Weight loss itself “Semi-starvation neurosis” Chance Can never exclude the possibility of chance Replication of findings across 9 of 13 studies Causal Biologically plausible ECS and CB1 receptor functions in CNS Increased incidence of depression in clinical trials Suicidality is a symptom of depression

Ascertainment Bias: 

Ascertainment Bias Patients who report common drug-related adverse events may be questioned more about other adverse events compared with placebo patients e.g., increased reporting of sexual dysfunction in depressed patients taking an antidepressant might lead to further questions about other adverse events and possibly increase the odds of reporting suicidal symptoms Most common adverse event in rimonabant-treated patients is nausea Would increased reporting of nausea in a population of non-depressed patients on rimonabant lead to increased reporting of suicidal symptoms? Depressive and anxiety events were reported on average 2 months later than nausea events

Ascertainment Bias (cont’d): 

Ascertainment Bias (cont’d) Increased contact of rimonabant-treated patients with investigators (for any reason) would provide more opportunity to voice suicidal symptoms compared with placebo-treated patients The mean and median number of study visits were 11.6 and 15 respectively, for both the placebo and rimonabant groups All datasets with unscheduled visits recorded were reviewed – no disproportionately between treatment groups in the number of unscheduled visits

Ascertainment Bias (cont’d): 

Ascertainment Bias (cont’d) Would have to be operative in 9 of 13 studies and explain odds ratios varying in magnitude from 1.4 to 16.7 Assumes an equal background rate of depression in placebo and rimonabant-treated subjects Recall: Depressed mood disorders: 9% for rimonabant subjects vs. 5% for placebo Discontinuation due to depressive disorders: 4.2% for rimonabant vs. 1.9% for placebo Required anti-depressant therapy: 5% for rimonabant vs. 3% for placebo Background rate should obscure differences over time

Rimonabant and SuicidalityWhat is the Nature of the Association?: 

Rimonabant and Suicidality What is the Nature of the Association? Bias Ascertainment Confounding Weight loss itself “Semi-starvation neurosis” Chance Can never exclude the possibility of chance Replication of findings across 9 of 13 studies Causal Biologically plausible ECS and CB1 receptor functions in CNS Increased incidence of depression in clinical trials Suicidality is a symptom of depression

Suicidality and Weight Change: 

Suicidality and Weight Change

Rimonabant and SuicidalityWhat is the Nature of the Association?: 

Rimonabant and Suicidality What is the Nature of the Association? Bias Ascertainment Confounding Weight loss itself “Semi-starvation neurosis” Chance Can never exclude the possibility of chance Replication of findings across 9 of 13 studies Causal Biologically plausible ECS and CB1 receptor functions in CNS Increased incidence of depression in clinical trials Suicidality is a symptom of depression

Rimonabant and SuicidalityWhat is the Nature of the Association?: 

Rimonabant and Suicidality What is the Nature of the Association? Bias Ascertainment Confounding Weight loss itself “Semi-starvation neurosis” Chance Can never exclude the possibility of chance Replication of findings across 9 of 13 studies Causal Biologically plausible ECS and CB1 receptor functions in CNS Increased incidence of depression in clinical trials Suicidality is a symptom of depression

Rimonabant and Suicidality Summary: 

Rimonabant and Suicidality Summary Meta-analysis indicates an increased risk for suicidality, specifically suicidal ideation, in subjects taking rimonabant 20 mg vs placebo ↑ Relative risk ~ 80 to 100% ↑ Absolute risk ~ 0.3% NNH ~ 1 in 300 patients Estimates may be low given that: Higher percentage of rimonabant patients dropped out of studies due to psychiatric adverse events

European Experience: 

European Experience Rimonabant approved in 30 countries As of March 1, 2007, an estimated 100,000 people (mostly from the UK and Germany) have been prescribed rimonabant Top 10 Preferred Terms: Depression Nausea Depressed mood Anxiety Fatigue Dizziness Sleep disorder Suicidal ideation Agitation Asthenia

EMEA PostMarketing Data(Cumulative Since Approval): 

EMEA PostMarketing Data (Cumulative Since Approval)

EMEA PostMarketing Data(Cumulative Since Approval): 

EMEA PostMarketing Data (Cumulative Since Approval) Cases of suicidal ideation: Orlistat (1998) 14 cases Sibutramine (1999) 15 cases Rimonabant (2006) 27 cases

Benefit vs. Harm: 

Benefit vs. Harm Weight loss TG, HDL HbA1C Fasting insulin Yet to be identified Psychiatric suicidality depression anxiety Neurological dizziness paresthesia/dysesthesia tremor seizure? Nausea/vomiting Yet to be identified Potential Benefits Potential Harms

CRESCENDO: 

CRESCENDO Comprehensive Rimonabant Evaluation Study of Cardiovascular Endpoints and Outcomes 17,000 abdominally obese patients at risk for cardiovascular disease Rimonabant 20 mg vs. placebo ~ 5 years in duration Primary outcome: myocardial infarction, stroke, or cardiovascular death To be completed in 2010

Acknowledgements: 

Acknowledgements Medical Review Team Pharmacology/Toxicology Eric Colman, M.D. John Gong, Ph.D. Mary Parks, M.D. Karen Davis-Bruno, Ph.D. Statistical Review Team Lee Ping Pian, Ph.D. J. Todd Sahlroot, Ph.D. Project Manager Patricia Madara

Questions for the Advisory Committee: 

Questions for the Advisory Committee 1. Please discuss your level of concern regarding rimonabant and psychiatric adverse events, in particular depression and suicidality, and neurological adverse events, in particular seizures, and the reasons behind your thinking on these issues. 2a. Do you believe that the currently available data sufficiently characterize rimonabant’s safety profile (vote requested)? 2b. If no, please discuss what additional data should be obtained.

Questions for the Advisory Committee: 

Questions for the Advisory Committee 3a. Based on the currently available data, do you believe rimonabant has a favorable risk-benefit profile and should be approved for the indication of weight management in individuals with a body mass index of > 30 kg/m2 and > 27 kg/m2 when accompanied by at least one comorbid condition (vote requested)? 3b. If no, please explain why and discuss what additional information the sponsor could obtain that might improve rimonabant’s risk-benefit profile.