- Email: [email protected]
A Retrospective, Pooled Suicidality Evaluation of Ropinirole Immediate Release and Controlled Release for the Treatment of Restless Legs Syndrome
Clinical Therapeutics/Volume 37, Number 5, 2015
A Retrospective, Pooled Suicidality Evaluation of Ropinirole Immediate Release and Controlled Release for the Treatment of Restless Legs Syndrome Clare W. Makumi, PharmD, MBA1; Walter Paska, PhD2; Katie Rolfe, MSc2; and Kenneth J. Shulman, DO1 1
Medicines Development Center, Neurosciences, GlaxoSmithKline, Research Triangle Park, North Carolina; and 2Medicines Development Center, Neurosciences, GlaxoSmithKline, Stockley Park, United Kingdom
ABSTRACT Purpose: Based on an internal request, GlaxoSmithKline conducted a retrospective pooled analysis of randomized controlled trials to compare suicidality in adult subjects with restless legs syndrome (RLS) who were being treated with ropinirole. The objective was to proactively evaluate the incidence of potentially suicidal thoughts or behaviors (suicidality) among patients with RLS treated with ropinirole immediate release (IR) or controlled release (CR). Methods: The US Food and Drug Administration approved methods previously used for the retrospective analysis of suicidality with antidepressants and anticonvulsants. Potential cases of suicidal thoughts and behavior were identified from searches of treatment-emergent adverse event preferred and verbatim terms; a review of serious adverse events; and searches of a priori–identified free text comment fields in the case report forms. Blinded case reports for these potential cases, in addition to all serious adverse events, were categorized by using the Columbia Classification Algorithm of Suicide Assessment. Findings: The dataset for this study comprised 1799 patients who received ropinirole (either formulation) and 1258 patients who received placebo. No signal for suicidality was detected for ropinirole in the treatment of patients with RLS. Implications: The pooled datasets in this study were not designed to prospectively assess for suicidal ideation or behavior. Any future studies in this area should include the collection of prespecified, detailed information regarding suicidality. (Clin Ther. 2015;37:1122–1127) & 2015 Elsevier HS Journals, Inc. All rights reserved.
1122
Key words: adverse event, C-CASA, pooled analysis, restless legs syndrome, retrospective, RLS.
INTRODUCTION Restless legs syndrome (RLS) is a neurologic sensorimotor disorder characterized by an urge to move the legs and is often accompanied or caused by uncomfortable sensations in the legs and (less frequently) the arms. The symptoms occur primarily at rest in the evening or at night and are at least partially and temporarily relieved during movement or activity.1,2 Several studies have shown that moderate to severe RLS affects 2% to 3% of adults (45 million US subjects).3 Up to 85% of patients also have involuntary, semi-rhythmic limb movements during sleep that are referred to as periodic limb movements of sleep. As a consequence of the sensory symptoms and/or periodic limb movements of sleep, RLS may have profound negative effects on sleep. Patients can initially experience insomnia or fatigue in addition to reduced concentration and memory, decreased motivation and drive, depression and anxiety, and reduced motor performance.2 Although an association has been shown between RLS symptoms and worsening mood compared with control populations, causality has not been directly determined.4 Patients with RLS reportedly have a higher incidence of depression and anxiety than matched control subjects.5 Depressed mood in RLS has been shown to be related to sleep disturbance. A study investigating the symptoms of major Accepted for publication February 8, 2015. http://dx.doi.org/10.1016/j.clinthera.2015.02.012 0149-2918/$ - see front matter & 2015 Elsevier HS Journals, Inc. All rights reserved.
Volume 37 Number 5
C.W. Makumi et al. depressive disorder in patients with RLS compared with patients without RLS found no difference in the major depressive disorder symptoms that were not associated with sleep impairment, including suicidal ideation.6 Suicide and suicidality-related events remain significant public health issues. Suicidality (suicidal thoughts and behavior) is associated with several risk factors, with depression and other mental health disorders being the primary risk factors.7 Although the pathogenesis of RLS is unknown, the involvement of the primary dopaminergic system is supported, in part, by the responsiveness of RLS patients to treatment with dopaminergic agents.5 Ropinirole, a nonergot D2/D3 agonist, is an approved treatment for RLS.8 During a routine internal GlaxoSmithKline (GSK) review of ropinirole immediate-release (IR) tablets indicated for the treatment of RLS, the determination was made to examine existing clinical trial data to proactively evaluate the risk of suicidality (suicidal ideation and suicidal behavior) in ropinirole-treated patients with RLS. The current article presents the results of that assessment.
MATERIALS AND METHODS Studies that met the following qualifications were included in the initial GSK assessment for suicidality: completed, randomized, double-blind, parallel-group, controlled studies (placebo or active comparator) with ropinirole (both IR and the controlled-release [CR] formulations) in subjects with RLS that were part of an existing internal integrated safety database. Table I provides a listing of all studies included in the initial pooled evaluation of suicidality. It was prespecified that if a single event of suicidality was identified in the initial assessment (ie, an event that was classified by a blinded GSK psychiatrist as completed suicide, suicide attempt, preparatory acts toward imminent suicidal behavior, suicide ideation), an additional review on an extended dataset would then be required. A single adverse event (AE) of suicidal behavior (a report of deliberate overdose) was identified during the initial assessment after the review of case report narratives, and the additional review was therefore performed. The additional studies included in this extended review met the following criteria: completed, randomized, double-blind, parallel-group, placebo-controlled studies with ropinirole (both IR and CR formulations) in patients with RLS. The following studies were excluded from this extended review: (1) three
May 2015
uncontrolled, open-label studies (studies 101468/206, 101468/192, and 101468/243); (2) one maintenance-ofeffect study in which all subjects received open-label ropinirole IR for 36 weeks before being randomized to receive ropinirole IR or placebo for an additional 12 weeks (study 101468/188); and (3) one study in which subjects were treated with ropinirole IR on an “as-needed” basis after a 3-week open-label phase (study RRO100310). Table II provides a listing of the additional studies included in the extended review of suicidality. The team used the same search strategy recommended by the US Food and Drug Administration for the evaluation of suicidality with antidepressants and anticonvulsants. In this search strategy, potential cases of suicidality were identified from: (1) searches of treatment-emergent AE preferred and verbatim terms; (2) review of all deaths and serious AEs; and (3) searches of free text comment fields in the study case report forms. Any free text string, or AEs coded to preferred terms, or verbatim terms that included the text string “-ACCID-”, “-INJUR-”, “-SUIC-”,“-OVERDOS-” , “-ACCIDENTAL OVERDOSE-”, “-ATTEMPT-”, “-CUT-”, “-GAS-”, “-HANG-”, “-HUNG-”, “-JUMP-”, “-MUTILAT-”, “-OVERDOS-”, “-SELF DAMAG-”, “-SELF HARM-”,“-SELF INFLICT-”, “-SHOOT-”, “-SLASH-”, “-SUIC-”, “-POISON-”, “-ASPHYX-”, “-SUFFOC-”, “-FIREARM-”, “-BURN-”, “-DROWN-”, “-GUN-”, “-IMMOLAT-”, “-MONOXIDE-”, “-TOX-”, “-LACERAT-”, “-DEATH-”, and “-DIE-” were identified as a potential case of suicidal behavior. A blinded listing of any data for patients who had a positive match in the aforementioned search was produced. All search results were sent to the clinical team for review; obvious false-positive findings (eg, “gas” for gastrointestinal) were flagged by the biostatistics and programming team. Subject numbering was different from the original study reports to ensure blinding of the data. The clinical team reviewed this listing and identified events that could potentially be related to suicidality risk factors or potential suicidal events. The events were identified as serious, such as lifethreatening events and hospitalizations; deaths; or overdoses, including those coded as “accidental.” A listing was produced, excluding any identified false-positive findings and including any already identified possible suicidality-related AEs from the search string.
1123
Clinical Therapeutics
Table I. Studies included in the suicidality assessment. Study 101468/190 101468/191 101468/194 101468/204 101468/205 101468/249
Treatment Ropinirole Placebo Ropinirole Placebo Ropinirole Placebo Ropinirole Ropinirole Ropinirole Placebo Ropinirole Placebo
No. of Subjects Exposed
Treatment Duration, wk
146 138 32 33 131 136 289 274 189 195 187 193
12
IR IR IR IR CR CR IR
12 12 12 12 12
IR ¼ immediate release; CR ¼ controlled release.
Case narratives were written for all events that were identified according to the clinical review as related to suicidality risk factors or potential suicidal events, with the exception of those that occurred pretreatment. Narratives were also provided for all serious AEs from the qualifying studies, irrespective of identification in the search term process. Narratives included information, where available, on: history of
suicidal thoughts, attempts or self-harm, specific number of days on study medication at time of event (not just dates), concurrent psychosocial stressors, psychiatric comorbidities, and family history of psychiatric disorders. A total of 181 blinded case narratives were sent to Columbia University for classification. The cases were classified according to the modified classification as described by Posner et al.9 Before
Table II. Studies included in the extended assessment of suicidality. Study ROP101892 RRL100013 RRL103660 ROR104836
Treatment Ropinirole Placebo Ropinirole Placebo Ropinirole Placebo Ropinirole Placebo
IR IR CR IR
No. of Subjects Exposed
Treatment Duration, wk
154* 149† 176 186 20 19 197 207
12 12 12 26
IR ¼ immediate release; CR ¼ controlled release. * A total of 158 subjects were randomized to receive ropinirole immediate release (IR) and are included in the analyses; however, only 154 received treatment. † A total of 151 subjects were randomized to receive placebo and are included in the analyses; however, only 149 received treatment.
1124
Volume 37 Number 5
C.W. Makumi et al. classification, all narratives were blinded to treatment, dates, and concomitant medications; given an identifier by Dr. Posner at Columbia (along with a GSK numeric identifier); and then delivered to the Columbia University group for classification. A spreadsheet was returned from the group containing the narrative identifiers and the corresponding classification ratings using the Columbia Classification Algorithm of Suicide Assessment (C-CASA) for mapping to original unblinded patient numbers before analysis. All cases were assigned 1 of the following categories from the C-CASA: (1) completed suicide; (2) suicide attempt; (3) preparatory acts toward imminent suicidal behavior; (4) suicidal ideation; (5) self-injurious behavior, intent unknown; (6) not enough information, fatal; (7) nonsuicidal self-injurious behavior; (8) other; and (9) not enough information, nonfatal. Items 1 through 3 inclusive were defined as “definitive suicidal behavior”; items 1 through 4 inclusive were defined as “definitive suicidal behavior and ideation.”
Analysis Methods Analyses were based on the classifications received from the Columbia University group. The primary objective was to compare the incidence of events occurring in patients receiving ropinirole (IR or CR) and placebo for definitive suicidal behavior and ideation or definitive suicidal behavior. The null hypothesis was that there was no difference between the 2 treatment groups; the alternative hypothesis was that a difference between the 2 treatment groups did exist. When the classifications were received from the Columbia University group, there were no identified cases of suicidal ideation. The incidence of events was therefore the same for definitive suicidal behavior and ideation and definitive suicidal behavior. Only 1 category was thus analyzed (that of definitive suicidal behavior); however, the results are applicable to either category. All analyses were based on the safety population (for all studies except ROP101892); the safety population was defined as all patients who were randomized to treatment and received at least 1 dose of trial medication. For study ROP101892, four patients from the ropinirole IR group and two patients from the placebo group were inadvertently included in the analysis, despite having received no treatment. The study conclusions were not influenced by the inclusion of these subjects.
May 2015
The incidence of events occurring during the ontreatment phase was compared between groups (ropinirole IR/CR combined and placebo). We had intended to adjust for study in the analyses; however, only 1 event occurred in study 101468/204, which did not include a placebo arm, and hence this adjustment was not appropriate. An unadjusted analysis using the exact method of StatXact (Cytel, Cambridge, Massaschusetts) was therefore performed. To assess the robustness of the analysis, a sensitivity analysis was performed by using Peto’s method.10
RESULTS Data were reviewed for 3057 patients (1799 receiving ropinirole and 1258 receiving placebo). A total of 181 cases were identified, and their blinded narratives sent to Columbia University for review. The original initial rating for the single subject classified as 2 (suicide attempt) was a tentative rating pending any additional information in the ropinirole group; this rating was subsequently amended to a score of 8 (other) after the Columbia University reviewer was provided with additional follow-up information regarding this subject’s event (details are provided in the Discussion). Of the 181 potential cases of suicidal behavior, all received a final Columbia rating of 8 (other), with the exception of 1 patient (details are provided in the Discussion) who was classified as 9 (not enough information, nonfatal). Therefore, on review of all available data, no case of either suicidal behavior or ideation was identified in either treatment group. The results from the original analysis indicated no statistically significant difference between the ropinirole (IR and CR) and placebo groups in the incidence of suicidal behavior (Columbia ratings 1–3). One of the 1799 ropinirole-treated patients (this subject was tentatively [pending additional information] classified under “suicide attempt”) and none of the 1258 patients in the placebo group reported suicidal behavior (exact method: odds ratio ¼ infinity [95% CI, 0.0–infinity], P ¼ 1.000; Peto’s method: odds ratio ¼ 5.5 [95% CI, 0.1–293.6], P ¼ 0.403). Because no events were classified as suicidal ideation (Columbia rating of 4), the results for the incidence of suicidal behavior or ideation are the same as those for suicidal behavior alone.
1125
Clinical Therapeutics
DISCUSSION Initially, 2 subjects received Columbia rating scores in which definitive suicidal behavior or ideation could not be ruled out based on the information provided. For 1 subject, a serious AE was reported as an “intentional overdose,” resulting in a tentative Columbia rating score of 2 (suicide attempt), pending any additional information. This event involved a 38-year-old female subject who was enrolled in a blinded study for the treatment of RLS. Six days after the start of the investigational product (ie, ropinirole IR), the subject experienced what was initially described as a deliberate drug overdose and was hospitalized. The subject had 4 empty bottles of study medication, but the quantity of medication ingested was unknown. No other event was reported at the time. Treatment with the investigational product was discontinued, and the subject was withdrawn from the study. The event resolved the next day, and the subject was discharged from the hospital. The investigator considered there was a reasonable possibility that the drug overdose may have been caused by the investigational product. At the request of the Columbia University reviewer, blinded medical records were obtained from the investigator. According to the records, the subject was treated in the emergency department. The treating physician indicated that the subject “took an overdose but did not believe it would kill her; the tablets are for ‘restless legs’ and she has been on them for one week. No suicide intent. No current thoughts of self-harm.” In the records, the subject was quoted as saying, “I just feel stupid. The tablets were part of a trial, and I thought that they could be the placebo.” The second subject was a 67-year-old woman who was randomized to the placebo group in a doubleblind, parallel-group study for the treatment of RLS. During treatment, the subject experienced mild lacerations on her forearms that were considered to be nonserious. The initial Columbia rating score was 9 (not enough information, nonfatal). Blinded medical records indicated that the lacerations occurred while she was doing yard work and were not associated with self-injurious behavior. For both of these subjects, the final Columbia rating was subsequently adjusted to a C-CASA score of 8 (other) based on the information in their medical records. On final review of all available data, the Columbia University group identified no cases of suicidal behavior or ideation in either treatment group.
1126
There are some limitations to this retrospective review of pooled data. It was conducted retrospectively from studies that were not designed to prospectively assess for suicidal ideation or behavior. The information was collected from case report forms, and these forms were not specifically designed to collect detailed information regarding suicidality.
CONCLUSIONS The total number of subjects for which data were reviewed was 3057 (1799 patients received ropinirole, and 1258 received placebo). A total of 181 potential cases of suicidal behavior were identified, with blinded narratives for these subjects’ case reports sent to Columbia University for review. All received a final C-CASA classification of 8 (other). No cases of suicidal behavior or ideation were identified according to this retrospective review, in either treatment group, on assessment of the complete information of subjects enrolled in trials of ropinirole (IR and CR) for the indication of RLS.
ACKNOWLEDGMENTS The authors thank Dr. Knezevic (GSK) for her contribution to the conduct and reporting of this study, as well as Andrea Jordan (ExecuPharm) for assistance with formating of the manuscript (funded by GSK).
CONFLICTS OF INTEREST This work was funded by GlaxoSmithKline Research & Development (study number 114236). Dr. Makumi and Ms. Rolfe are employees of, and stockholders in, GSK. Dr. Paska (retired) and Dr. Shulman (now at Genzyme Sanofi) are former GSK employees; Dr. Paska remains a stockholder in GSK. The authors have indicated that they have no other conflicts of interest regarding the content of this article. All authors contributed to the protocol development and design, data analysis and data interpretation, in addition to authoring and review.
REFERENCES 1. Allen RP, Picchietti D, Hening WA, et al. Restless legs syndrome: diagnostic criteria, special considerations, and epidemiology. A report from the restless legs syndrome diagnosis and epidemiology workshop at the National Institutes of Health. Sleep Med. 2003;4:101–119.
Volume 37 Number 5
C.W. Makumi et al. 2. Earley CJ. Restless legs syndrome. N Engl J Med. 2003;348:2103–2109. 3. http://www.ninds.nih.gov/disorders/restless_legs/detail_restless_ legs.html. 4. Picchietti D, Winkelman JW. Restless legs syndrome, periodic limb movements in sleep, and depression. Sleep. 2005;28:891–898. 5. Becker PM. The biopsychosocial effects of restless legs syndrome (RLS). Neuropsychiatr Dis Treat. 2006; 2:505–512. 6. Winkelman JW, Plant DT. Foundations of Psychiatric Medicine, Chapter 8: Sleep-related movement disorders. 2010: 138–139. 7. Moscicki EK. Epidemiology of completed and attempted suicide: toward a framework for prevention. Clin Neurosci Res. 2001;1:310–323. 8. Product Information, REQUIP (ropinirole tablets) Product Information, April 2009. #8. 9. Posner K, Oquendo MA, Gould M, et al. Columbia Classification Algorithm of Suicide Assessment (C-CASA): classification of suicidal events in the FDA’s pediatric suicidal risk analysis of antidepressants. Am J Psychiatry. 2007;174:1035–1043. 10. Yusuf S, Peto R, Lewis J, Colins R, Sleight P. Beta blockade during and after myocardial infarction. An overview of randomized trials. Prog Cardiovasc Dis. 1985;27:335–371.
Address correspondence to: Clare W. Makumi, PharmD, MBA, Medical Services, North America Medical Affairs, GlaxoSmithKline, 5 Moore Drive, PO Box 13398, Research Triangle Park, NC 27709. E-mail: clare. [email protected]
May 2015
1127
A Retrospective, Pooled Suicidality Evaluation of Ropinirole Immediate Release and Controlled Release for the Treatment of Restless Legs Syndrome Clare W. Makumi, PharmD, MBA1; Walter Paska, PhD2; Katie Rolfe, MSc2; and Kenneth J. Shulman, DO1 1
Medicines Development Center, Neurosciences, GlaxoSmithKline, Research Triangle Park, North Carolina; and 2Medicines Development Center, Neurosciences, GlaxoSmithKline, Stockley Park, United Kingdom
ABSTRACT Purpose: Based on an internal request, GlaxoSmithKline conducted a retrospective pooled analysis of randomized controlled trials to compare suicidality in adult subjects with restless legs syndrome (RLS) who were being treated with ropinirole. The objective was to proactively evaluate the incidence of potentially suicidal thoughts or behaviors (suicidality) among patients with RLS treated with ropinirole immediate release (IR) or controlled release (CR). Methods: The US Food and Drug Administration approved methods previously used for the retrospective analysis of suicidality with antidepressants and anticonvulsants. Potential cases of suicidal thoughts and behavior were identified from searches of treatment-emergent adverse event preferred and verbatim terms; a review of serious adverse events; and searches of a priori–identified free text comment fields in the case report forms. Blinded case reports for these potential cases, in addition to all serious adverse events, were categorized by using the Columbia Classification Algorithm of Suicide Assessment. Findings: The dataset for this study comprised 1799 patients who received ropinirole (either formulation) and 1258 patients who received placebo. No signal for suicidality was detected for ropinirole in the treatment of patients with RLS. Implications: The pooled datasets in this study were not designed to prospectively assess for suicidal ideation or behavior. Any future studies in this area should include the collection of prespecified, detailed information regarding suicidality. (Clin Ther. 2015;37:1122–1127) & 2015 Elsevier HS Journals, Inc. All rights reserved.
1122
Key words: adverse event, C-CASA, pooled analysis, restless legs syndrome, retrospective, RLS.
INTRODUCTION Restless legs syndrome (RLS) is a neurologic sensorimotor disorder characterized by an urge to move the legs and is often accompanied or caused by uncomfortable sensations in the legs and (less frequently) the arms. The symptoms occur primarily at rest in the evening or at night and are at least partially and temporarily relieved during movement or activity.1,2 Several studies have shown that moderate to severe RLS affects 2% to 3% of adults (45 million US subjects).3 Up to 85% of patients also have involuntary, semi-rhythmic limb movements during sleep that are referred to as periodic limb movements of sleep. As a consequence of the sensory symptoms and/or periodic limb movements of sleep, RLS may have profound negative effects on sleep. Patients can initially experience insomnia or fatigue in addition to reduced concentration and memory, decreased motivation and drive, depression and anxiety, and reduced motor performance.2 Although an association has been shown between RLS symptoms and worsening mood compared with control populations, causality has not been directly determined.4 Patients with RLS reportedly have a higher incidence of depression and anxiety than matched control subjects.5 Depressed mood in RLS has been shown to be related to sleep disturbance. A study investigating the symptoms of major Accepted for publication February 8, 2015. http://dx.doi.org/10.1016/j.clinthera.2015.02.012 0149-2918/$ - see front matter & 2015 Elsevier HS Journals, Inc. All rights reserved.
Volume 37 Number 5
C.W. Makumi et al. depressive disorder in patients with RLS compared with patients without RLS found no difference in the major depressive disorder symptoms that were not associated with sleep impairment, including suicidal ideation.6 Suicide and suicidality-related events remain significant public health issues. Suicidality (suicidal thoughts and behavior) is associated with several risk factors, with depression and other mental health disorders being the primary risk factors.7 Although the pathogenesis of RLS is unknown, the involvement of the primary dopaminergic system is supported, in part, by the responsiveness of RLS patients to treatment with dopaminergic agents.5 Ropinirole, a nonergot D2/D3 agonist, is an approved treatment for RLS.8 During a routine internal GlaxoSmithKline (GSK) review of ropinirole immediate-release (IR) tablets indicated for the treatment of RLS, the determination was made to examine existing clinical trial data to proactively evaluate the risk of suicidality (suicidal ideation and suicidal behavior) in ropinirole-treated patients with RLS. The current article presents the results of that assessment.
MATERIALS AND METHODS Studies that met the following qualifications were included in the initial GSK assessment for suicidality: completed, randomized, double-blind, parallel-group, controlled studies (placebo or active comparator) with ropinirole (both IR and the controlled-release [CR] formulations) in subjects with RLS that were part of an existing internal integrated safety database. Table I provides a listing of all studies included in the initial pooled evaluation of suicidality. It was prespecified that if a single event of suicidality was identified in the initial assessment (ie, an event that was classified by a blinded GSK psychiatrist as completed suicide, suicide attempt, preparatory acts toward imminent suicidal behavior, suicide ideation), an additional review on an extended dataset would then be required. A single adverse event (AE) of suicidal behavior (a report of deliberate overdose) was identified during the initial assessment after the review of case report narratives, and the additional review was therefore performed. The additional studies included in this extended review met the following criteria: completed, randomized, double-blind, parallel-group, placebo-controlled studies with ropinirole (both IR and CR formulations) in patients with RLS. The following studies were excluded from this extended review: (1) three
May 2015
uncontrolled, open-label studies (studies 101468/206, 101468/192, and 101468/243); (2) one maintenance-ofeffect study in which all subjects received open-label ropinirole IR for 36 weeks before being randomized to receive ropinirole IR or placebo for an additional 12 weeks (study 101468/188); and (3) one study in which subjects were treated with ropinirole IR on an “as-needed” basis after a 3-week open-label phase (study RRO100310). Table II provides a listing of the additional studies included in the extended review of suicidality. The team used the same search strategy recommended by the US Food and Drug Administration for the evaluation of suicidality with antidepressants and anticonvulsants. In this search strategy, potential cases of suicidality were identified from: (1) searches of treatment-emergent AE preferred and verbatim terms; (2) review of all deaths and serious AEs; and (3) searches of free text comment fields in the study case report forms. Any free text string, or AEs coded to preferred terms, or verbatim terms that included the text string “-ACCID-”, “-INJUR-”, “-SUIC-”,“-OVERDOS-” , “-ACCIDENTAL OVERDOSE-”, “-ATTEMPT-”, “-CUT-”, “-GAS-”, “-HANG-”, “-HUNG-”, “-JUMP-”, “-MUTILAT-”, “-OVERDOS-”, “-SELF DAMAG-”, “-SELF HARM-”,“-SELF INFLICT-”, “-SHOOT-”, “-SLASH-”, “-SUIC-”, “-POISON-”, “-ASPHYX-”, “-SUFFOC-”, “-FIREARM-”, “-BURN-”, “-DROWN-”, “-GUN-”, “-IMMOLAT-”, “-MONOXIDE-”, “-TOX-”, “-LACERAT-”, “-DEATH-”, and “-DIE-” were identified as a potential case of suicidal behavior. A blinded listing of any data for patients who had a positive match in the aforementioned search was produced. All search results were sent to the clinical team for review; obvious false-positive findings (eg, “gas” for gastrointestinal) were flagged by the biostatistics and programming team. Subject numbering was different from the original study reports to ensure blinding of the data. The clinical team reviewed this listing and identified events that could potentially be related to suicidality risk factors or potential suicidal events. The events were identified as serious, such as lifethreatening events and hospitalizations; deaths; or overdoses, including those coded as “accidental.” A listing was produced, excluding any identified false-positive findings and including any already identified possible suicidality-related AEs from the search string.
1123
Clinical Therapeutics
Table I. Studies included in the suicidality assessment. Study 101468/190 101468/191 101468/194 101468/204 101468/205 101468/249
Treatment Ropinirole Placebo Ropinirole Placebo Ropinirole Placebo Ropinirole Ropinirole Ropinirole Placebo Ropinirole Placebo
No. of Subjects Exposed
Treatment Duration, wk
146 138 32 33 131 136 289 274 189 195 187 193
12
IR IR IR IR CR CR IR
12 12 12 12 12
IR ¼ immediate release; CR ¼ controlled release.
Case narratives were written for all events that were identified according to the clinical review as related to suicidality risk factors or potential suicidal events, with the exception of those that occurred pretreatment. Narratives were also provided for all serious AEs from the qualifying studies, irrespective of identification in the search term process. Narratives included information, where available, on: history of
suicidal thoughts, attempts or self-harm, specific number of days on study medication at time of event (not just dates), concurrent psychosocial stressors, psychiatric comorbidities, and family history of psychiatric disorders. A total of 181 blinded case narratives were sent to Columbia University for classification. The cases were classified according to the modified classification as described by Posner et al.9 Before
Table II. Studies included in the extended assessment of suicidality. Study ROP101892 RRL100013 RRL103660 ROR104836
Treatment Ropinirole Placebo Ropinirole Placebo Ropinirole Placebo Ropinirole Placebo
IR IR CR IR
No. of Subjects Exposed
Treatment Duration, wk
154* 149† 176 186 20 19 197 207
12 12 12 26
IR ¼ immediate release; CR ¼ controlled release. * A total of 158 subjects were randomized to receive ropinirole immediate release (IR) and are included in the analyses; however, only 154 received treatment. † A total of 151 subjects were randomized to receive placebo and are included in the analyses; however, only 149 received treatment.
1124
Volume 37 Number 5
C.W. Makumi et al. classification, all narratives were blinded to treatment, dates, and concomitant medications; given an identifier by Dr. Posner at Columbia (along with a GSK numeric identifier); and then delivered to the Columbia University group for classification. A spreadsheet was returned from the group containing the narrative identifiers and the corresponding classification ratings using the Columbia Classification Algorithm of Suicide Assessment (C-CASA) for mapping to original unblinded patient numbers before analysis. All cases were assigned 1 of the following categories from the C-CASA: (1) completed suicide; (2) suicide attempt; (3) preparatory acts toward imminent suicidal behavior; (4) suicidal ideation; (5) self-injurious behavior, intent unknown; (6) not enough information, fatal; (7) nonsuicidal self-injurious behavior; (8) other; and (9) not enough information, nonfatal. Items 1 through 3 inclusive were defined as “definitive suicidal behavior”; items 1 through 4 inclusive were defined as “definitive suicidal behavior and ideation.”
Analysis Methods Analyses were based on the classifications received from the Columbia University group. The primary objective was to compare the incidence of events occurring in patients receiving ropinirole (IR or CR) and placebo for definitive suicidal behavior and ideation or definitive suicidal behavior. The null hypothesis was that there was no difference between the 2 treatment groups; the alternative hypothesis was that a difference between the 2 treatment groups did exist. When the classifications were received from the Columbia University group, there were no identified cases of suicidal ideation. The incidence of events was therefore the same for definitive suicidal behavior and ideation and definitive suicidal behavior. Only 1 category was thus analyzed (that of definitive suicidal behavior); however, the results are applicable to either category. All analyses were based on the safety population (for all studies except ROP101892); the safety population was defined as all patients who were randomized to treatment and received at least 1 dose of trial medication. For study ROP101892, four patients from the ropinirole IR group and two patients from the placebo group were inadvertently included in the analysis, despite having received no treatment. The study conclusions were not influenced by the inclusion of these subjects.
May 2015
The incidence of events occurring during the ontreatment phase was compared between groups (ropinirole IR/CR combined and placebo). We had intended to adjust for study in the analyses; however, only 1 event occurred in study 101468/204, which did not include a placebo arm, and hence this adjustment was not appropriate. An unadjusted analysis using the exact method of StatXact (Cytel, Cambridge, Massaschusetts) was therefore performed. To assess the robustness of the analysis, a sensitivity analysis was performed by using Peto’s method.10
RESULTS Data were reviewed for 3057 patients (1799 receiving ropinirole and 1258 receiving placebo). A total of 181 cases were identified, and their blinded narratives sent to Columbia University for review. The original initial rating for the single subject classified as 2 (suicide attempt) was a tentative rating pending any additional information in the ropinirole group; this rating was subsequently amended to a score of 8 (other) after the Columbia University reviewer was provided with additional follow-up information regarding this subject’s event (details are provided in the Discussion). Of the 181 potential cases of suicidal behavior, all received a final Columbia rating of 8 (other), with the exception of 1 patient (details are provided in the Discussion) who was classified as 9 (not enough information, nonfatal). Therefore, on review of all available data, no case of either suicidal behavior or ideation was identified in either treatment group. The results from the original analysis indicated no statistically significant difference between the ropinirole (IR and CR) and placebo groups in the incidence of suicidal behavior (Columbia ratings 1–3). One of the 1799 ropinirole-treated patients (this subject was tentatively [pending additional information] classified under “suicide attempt”) and none of the 1258 patients in the placebo group reported suicidal behavior (exact method: odds ratio ¼ infinity [95% CI, 0.0–infinity], P ¼ 1.000; Peto’s method: odds ratio ¼ 5.5 [95% CI, 0.1–293.6], P ¼ 0.403). Because no events were classified as suicidal ideation (Columbia rating of 4), the results for the incidence of suicidal behavior or ideation are the same as those for suicidal behavior alone.
1125
Clinical Therapeutics
DISCUSSION Initially, 2 subjects received Columbia rating scores in which definitive suicidal behavior or ideation could not be ruled out based on the information provided. For 1 subject, a serious AE was reported as an “intentional overdose,” resulting in a tentative Columbia rating score of 2 (suicide attempt), pending any additional information. This event involved a 38-year-old female subject who was enrolled in a blinded study for the treatment of RLS. Six days after the start of the investigational product (ie, ropinirole IR), the subject experienced what was initially described as a deliberate drug overdose and was hospitalized. The subject had 4 empty bottles of study medication, but the quantity of medication ingested was unknown. No other event was reported at the time. Treatment with the investigational product was discontinued, and the subject was withdrawn from the study. The event resolved the next day, and the subject was discharged from the hospital. The investigator considered there was a reasonable possibility that the drug overdose may have been caused by the investigational product. At the request of the Columbia University reviewer, blinded medical records were obtained from the investigator. According to the records, the subject was treated in the emergency department. The treating physician indicated that the subject “took an overdose but did not believe it would kill her; the tablets are for ‘restless legs’ and she has been on them for one week. No suicide intent. No current thoughts of self-harm.” In the records, the subject was quoted as saying, “I just feel stupid. The tablets were part of a trial, and I thought that they could be the placebo.” The second subject was a 67-year-old woman who was randomized to the placebo group in a doubleblind, parallel-group study for the treatment of RLS. During treatment, the subject experienced mild lacerations on her forearms that were considered to be nonserious. The initial Columbia rating score was 9 (not enough information, nonfatal). Blinded medical records indicated that the lacerations occurred while she was doing yard work and were not associated with self-injurious behavior. For both of these subjects, the final Columbia rating was subsequently adjusted to a C-CASA score of 8 (other) based on the information in their medical records. On final review of all available data, the Columbia University group identified no cases of suicidal behavior or ideation in either treatment group.
1126
There are some limitations to this retrospective review of pooled data. It was conducted retrospectively from studies that were not designed to prospectively assess for suicidal ideation or behavior. The information was collected from case report forms, and these forms were not specifically designed to collect detailed information regarding suicidality.
CONCLUSIONS The total number of subjects for which data were reviewed was 3057 (1799 patients received ropinirole, and 1258 received placebo). A total of 181 potential cases of suicidal behavior were identified, with blinded narratives for these subjects’ case reports sent to Columbia University for review. All received a final C-CASA classification of 8 (other). No cases of suicidal behavior or ideation were identified according to this retrospective review, in either treatment group, on assessment of the complete information of subjects enrolled in trials of ropinirole (IR and CR) for the indication of RLS.
ACKNOWLEDGMENTS The authors thank Dr. Knezevic (GSK) for her contribution to the conduct and reporting of this study, as well as Andrea Jordan (ExecuPharm) for assistance with formating of the manuscript (funded by GSK).
CONFLICTS OF INTEREST This work was funded by GlaxoSmithKline Research & Development (study number 114236). Dr. Makumi and Ms. Rolfe are employees of, and stockholders in, GSK. Dr. Paska (retired) and Dr. Shulman (now at Genzyme Sanofi) are former GSK employees; Dr. Paska remains a stockholder in GSK. The authors have indicated that they have no other conflicts of interest regarding the content of this article. All authors contributed to the protocol development and design, data analysis and data interpretation, in addition to authoring and review.
REFERENCES 1. Allen RP, Picchietti D, Hening WA, et al. Restless legs syndrome: diagnostic criteria, special considerations, and epidemiology. A report from the restless legs syndrome diagnosis and epidemiology workshop at the National Institutes of Health. Sleep Med. 2003;4:101–119.
Volume 37 Number 5
C.W. Makumi et al. 2. Earley CJ. Restless legs syndrome. N Engl J Med. 2003;348:2103–2109. 3. http://www.ninds.nih.gov/disorders/restless_legs/detail_restless_ legs.html. 4. Picchietti D, Winkelman JW. Restless legs syndrome, periodic limb movements in sleep, and depression. Sleep. 2005;28:891–898. 5. Becker PM. The biopsychosocial effects of restless legs syndrome (RLS). Neuropsychiatr Dis Treat. 2006; 2:505–512. 6. Winkelman JW, Plant DT. Foundations of Psychiatric Medicine, Chapter 8: Sleep-related movement disorders. 2010: 138–139. 7. Moscicki EK. Epidemiology of completed and attempted suicide: toward a framework for prevention. Clin Neurosci Res. 2001;1:310–323. 8. Product Information, REQUIP (ropinirole tablets) Product Information, April 2009. #8. 9. Posner K, Oquendo MA, Gould M, et al. Columbia Classification Algorithm of Suicide Assessment (C-CASA): classification of suicidal events in the FDA’s pediatric suicidal risk analysis of antidepressants. Am J Psychiatry. 2007;174:1035–1043. 10. Yusuf S, Peto R, Lewis J, Colins R, Sleight P. Beta blockade during and after myocardial infarction. An overview of randomized trials. Prog Cardiovasc Dis. 1985;27:335–371.
Address correspondence to: Clare W. Makumi, PharmD, MBA, Medical Services, North America Medical Affairs, GlaxoSmithKline, 5 Moore Drive, PO Box 13398, Research Triangle Park, NC 27709. E-mail: clare. [email protected]
May 2015
1127