Rimonabant: safety issues

Rimonabant ( Acomplia / Zimulti ) is a cannabinoid receptor antagonist. It acts by blocking a specific type of receptor called cannabinoid type 1 ( CB1 ) receptors. These receptors are found in the nervous system and are part of the system that the body uses to control food intake. By blocking the receptors, Rimonabant can help patients to reduce food intake and to lose weight. The receptors are also found in peripheral tissues including adipocytes.

Review of the preclinical and clinical data has raised concern about associations between Rimonabant and increased frequencies of psychiatric adverse events, including suicidality, an ill-defined constellation of neurological signs and symptoms, and seizures.

A. Psychiatric adverse events

Endocannabinoids are important modulators in pathological conditions such as anxiety, phobias, depression, and posttraumatic stress disorders.
Therefore, the emergence of psychiatric symptoms with the use of a cannabinoid receptor Antagonist / inverse agonist is biologically plausible.

In the pooled RIO studies, for subjects receiving the same treatment during the whole study, 26% of Rimonabant 20mg treated subjects versus 14% of placebo treated subjects experienced a psychiatric symptom reported as an adverse event. Specifically, 9% of Rimonabant 20mg treated subjects versus 5% of placebo treated subjects reported symptoms of depression ( depressed mood; depression; depressive symptom; or major depression ).

A retrospective analysis of source documentation was performed by Sanofi-Aventis in order to obtain additional data on specific psychiatric events. Attempts were made to capture associated psychiatric symptoms, most notably psychomotor agitation, psychomotor retardation, anxiety, suicidal ideation, aggressivity, irritability, etc.
An individual could report multiple symptoms; these symptoms were not recorded on an adverse event report form and are therefore over and above the total number of adverse events noted above.
These included 155 additional symptoms associated with depressed mood disorders and 208 additional symptoms of anxiety disorders in subjects receiving Rimonabant 20mg versus 49 and 51, respectively in the placebo group.
Overall, the number of associated symptoms reported by subjects reporting a psychiatric adverse event was 427 in the Rimonabant 20mg treated group and 118 in the placebo-treated group. 24.
The relative risk for psychiatric adverse events in the Rimonabant 20mg versus placebo groups ranged from 1.5 to 2.5 in the four RIO studies. When considered in aggregate, the overall relative risk for psychiatric adverse events in the Rimonabant 20mg versus placebo group was 1.9.

These psychiatric adverse events more often necessitated discontinuation of study drug; more often required concomitant treatment ( pharmacologic and/or psychotherapy ); and were more often reported as “not recovered” or “recovering” at end of study, in the Rimonabant 20mg treated group versus the placebo-treated group.
The number of subjects requiring the institution of an anxiolytic or hypnotic agent for a psychiatric adverse event was: 185 subjects ( 8.5% ) on Rimonabant 20mg versus 102 subjects ( 4.6% ) on Rimonabant 5mg, and 66 subjects ( 4.1% ) on placebo. Another 104 subjects ( 4.8% ) on Rimonabant 20mg versus 88 subjects ( 4.0% ) on Rimonabant 5mg and 46 subjects ( 2.9% ) on placebo required the institution of an antidepressant agent for a psychiatric adverse event.


To investigate a signal for suicidality detected during review of the original NDA submission, DMEP requested that Sanofi-Aventis obtain a formal assessment of suicidality from Kelly Posner’s group at Columbia University.
Posner and her colleagues have been integrally involved in the recent assessment by FDA of suicidality in patients taking antidepressant drugs. The Columbia University group’s method of assessment is based on a blinded classification of cases.
Sanofi-Aventis searched the original clinical adverse event database to identify patients for whom additional information was needed. For these events, efforts were made to better document the case, either from source documents already collected during the course of the studies, or collected after returning to the sites.
Patient narratives were then prepared or updated and submitted to Posner’s group.

A total of 1201 patient-narratives were assessed in a strictly blinded manner by the Columbia University group. Ninety-one ( 91 ) cases were classified as either possibly ( Columbia categories 5, 6, or 9 ), or definitely ( Columbia categories 1, 2, 3, or 4 ) suicidal; this includes 5 cases which occurred on Haloperidol active treatment.

A total of 13 studies were used in the analyses: Study ACT4389 ( which had no 20 mg Rimonabant treatment group ) and study EFC4798 ( which had no placebo treatment group ) and study DRI5747 ( which did not have a clinical study report completed as of the cut-off date ) were excluded from the analyses. Studies EFC4743 and EFC4796 re-randomized patients during a maintenance phase treatment after the first randomized treatment. Only data from the first randomization were used in the analyses.

Thus, the total number of suicidality cases contributing to the analyses is 74 ( 20 on placebo, 8 on Rimonabant 5mg, and 46 on Rimonabant 20mg ).
The overall odds ratio for the incidence of suicidality: 20mg versus placebo for the cases indicated above was 1.9.
When limited to the 7 obesity studies, the odds ratio for incidence of suicidality: 20mg versus placebo was 1.8.

Roughly 50% of the subjects in the Rimonabant and placebo groups withdrew early from the trials, with more Rimonabant subjects doing so due to depression, anxiety, mood alteration with depressive symptoms, and the need for antidepressant medication. Given the lack of systematic follow-up of these subjects and Rimonabant’s long half-life ( ~16 days on average ), the results of the above analyses should be viewed as incomplete at best and at worse as an underestimate of Rimonabant’s risk for suicidality.

According to Sanofi-Aventis, in ongoing trials as of the December 18, 2006 cut-off date, data were available on 17 unblinded cases of suicidality - 11 on Rimonabant 20mg and 6 on placebo. The Rimonabant 20 mg cases included 1 completed suicide, 1 self-injurious ideation, 8 suicidal ideations, and 1 depression suicidal; the placebo cases included 2 suicide attempts and 5 suicidal ideations.
It should be noted that the FDA had also received 2 additional reports during this time period - one of “homicidal ideation” in a subject receiving Rimonabant 20mg in study PMC_0172 and one of suicide attempt in a subject receiving Rimonabant 20 mg in study EFC5823.
Subsequent to the sponsor’s submission of the safety update in March 2007, the following reports of suicidality have been received: 2 reports of suicide attempt - one in a 60-year-old female randomized to 20mg Rimonabant in the CRESCENDO trial and one in a 56-year-old male randomised to Rimonabant 20mg in the CRESCENDO trial; a report of a suicide gesture in a 37-year-old female randomized to 20mg Rimonabant in the RAPSODI trial; and a report of suicidal ideation in a 64-year-old female randomized to 20 mg Rimonabant in the CRESCENDO trial.

In the entire Rimonabant clinical trial database, there have been 2 completed suicides – one in RIO North America in a subject taking Rimonabant 5mg and one in the ongoing study STRADIVARIUS in a subject taking Rimonabant 20mg.

B. Concurrent use of antiobesity and antidepressant medication

All patients who were placed on antidepressant therapy were to be discontinued from the RIO studies. This was done to avoid confounding the weight-loss data.

Given the increased incidence of depression-related adverse events in subjects treated with Rimonabant 20mg versus placebo and the lack of data on the efficacy and safety of concomitant use of Rimonabant with antidepressants, the FDA obtained concurrency prescription-use data for antiobesity and antidepressant medication.

During the time period covering 2004 through 2006, roughly 580,000 raw patients per year received a prescription for one of the following weight-loss drugs: Phentermine ( Adipex-P ), Orlistat ( Xenical ), Sibutramine ( Meridia, Reductil ), or Diethylpropion ( Tenuate ).

Approximately 30% of these raw patients received a concurrent prescription for an antidepressant medication.

C. Neurological adverse events

CB1 receptor density is particularly high in the cerebellum, cortex, hippocampus, hypothalamus, and basal ganglia – areas of the brain that affect memory, motor function, and reward behaviors. They are also present on the peripheral nerves where they play a neuroprotective role.

Neurological symptoms, including sensory changes, motor impairments, and cognitive difficulties appeared commonly in the clinical trials, but were not well characterized or evaluated in detail.

Sanofi-Aventis was asked to provide additional data from ongoing and/or future trials in which appropriate attention was given to capturing and following-up on treatment-emergent neurological symptoms.

Neurological symptoms, while vague, occurred with greater frequency in Rimonabant 20mg treated patients than in placebo patients: 27.4% and 24.4% respectively. Among subjects treated with Rimonabant 20mg, the most commonly reported neurological symptoms were: headache ( 10% ), dizziness ( 8.6% ), and paresthesia /hypoaesthesia / dysaesthesia ( 3.3% ); while among subjects treated with placebo, the most commonly reported neurological symptoms were: headache ( 12.7% ), dizziness ( 5.6% ), paresthesia / hypoaesthesia / dysaesthesia ( 2.1% ).
Dizziness and vertigo occurred with greater frequency in the Rimonabant 20mg group than in the placebo group – 9.6% and 6.1%, respectively. Motor impairment occurred with greater frequency in the Rimonabant 20mg group than in the placebo group – 1.7% and 0.12%, respectively – and was driven predominantly by “tremor” and “balance disorder”.
Cognitive disorders occurred with greater frequency in the Rimonabant 20mg group than in the placebo group – 3.5% and 2.0%, respectively – and were driven predominantly by “mental impairment”, “somnolence”, and “disturbance in thinking / perception”.

These neurological adverse events may well have contributed to the disproportionate number of subjects who sustained injuries ( contusions, concussions, falls, road traffic accidents, whiplash, and injuries ) during the RIO trials in the Rimonabant 20mg group ( 6.9% ) vs. the placebo group ( 3.8% ).

Paresthesia, dysaesthesia, and hypoaesthesia occurred with slightly greater frequency in the Rimonabant 20mg group than in the placebo group – 3.3% and 2.1%, respectively.
However, when the studies in diabetic patients were analyzed separately, the occurrence of these symptoms was much greater in the Rimonabant 20mg group versus placebo.
In RIO-Diabetes and in SERENADE ( conducted in treatment-naïve Type 2 diabetics ) – approximately 5% of Rimonabant 20mg treated subjects versus 1.2% of placebo treated patients experienced paresthesia, dysaesthesia, or hypoaesthesia.
In RIO Diabetes and SERENADE, the overall relative risk for the incidence of a neurological adverse event for Rimonabant 20mg versus placebo was 3.1.

When multiple sclerosis is induced by viral inoculation in CB1 knockout mice or in mice treated with a CB1 receptor antagonist the neurodegenerative process is more severe. This suggests that CB1 antagonism may exacerbate inflammatory demyelinating diseases in humans.
Five cases of confirmed multiple sclerosis ( 3 ) or suspicion of demyelinating disease ( 2 ) have been reported from Rimonabant trials as of 18 December 2006. Of these, 1 ( 0.05% ) was in a patient on placebo; 2 were in patients on Rimonabant 5mg ( 0.09% ); and 2 were in patients on Rimonabant 20mg ( 0.05% ). The 2 subjects on Rimonabant 20mg who were suspected of having multiple sclerosis were both from smoking cessation trials ( 1 of them had a medical history of multiple sclerosis, but experienced exacerbation of her symptoms on Rimonabant ); the other 3 subjects were from obesity trials. One of the cases of multiple sclerosis from the obesity trials was published as a case report. This case was notable because recovery to near normal was noted within weeks after discontinuation of Rimonabant treatment.

D. Seizures

Cannabinoids possess anticonvulsant properties and the endocannabinoid system has been implicated in regulating seizure duration and frequency. It is speculated that epileptiform seizure activity elicits an increase in the “on-demand” synthesis of endocannabinoids resulting in increased activation of presynaptic CB1 receptors with subsequent regulation of neuronal hyperexcitability and seizure termination.

In animals, Rimonabant accumulates in the brain with multiple dosing, therefore AUC/Cmax ratios probably over-estimate safety margins in humans.
In preclinical evaluations, approximately 6% of rats and mice and 20% ( 2/10 ) of monkeys developed seizures while receiving long-term treatment with doses of Rimonabant 0.5-2 times the 20mg dose proposed for marketing. Approximately 1.5% of control mice developed seizures, while none of the control rats or monkeys did so.

Nineteen cases of seizure were reported in the completed Rimonabant clinical trials. Of these, three were excluded from the analyses – 2 cases that occurred during placebo run-in and one case that occurred > 3 months after dosing. Of the remaining sixteen, eleven were adjudicated as “likely” or “possible” by 2 independent neurologists: 6 cases in Rimonabant 20mg groups ( 5 in obesity trials; and 1 in a smoking cessation trial ), 2 cases in the Rimonabant 5mg groups ( both in obesity trials ), and 3 cases in the placebo groups ( 1 in an obesity trial; 1 in a smoking cessation trial; and 1 in the schizophrenia trial ).

Source: FDA, 2007


Link: Xapedia - Medical Encyclopedia