Investigation into the natural course of epilepsy and the implications on clinical trials

Proposal
1335

Title of Proposed Research

Lead Researcher

Daniel Goldenholz

Affiliation

NINDS, National Institutes of Health

Funding Source

This research is funded by the National Institute of Neurological Disorders and Stroke (NINDS) NIH Division of Intramural Research.

Potential Conflicts of Interest

None

Data Sharing Agreement Date

06 December 2016

Lay Summary

BACKGROUND
Epilepsy affects at least 2.2 million Americans, costs $9.5-$12.5 billion annually, and has a 10-fold increased risk of sudden death compared to the general public. Medicines currently do not help 1 out of 3 patients with this disease, and surgery either is ineffective or unavailable to many of these patients. When new medicines are tested in clinical studies, they are compared to placebo. If the medicine is stronger than placebo, then it is considered effective.

Clinical studies of epilepsy report a “placebo effect” of 4-19%. Studies in the natural course of epilepsy have found 30-60% of patients that become seizure-free from their disease temporarily or permanently. Recently a simulation using at the largest database of seizures found effects identical to the “placebo effect” without any placebo being given. Taken together, it seems possible that the “placebo effect” of clinical studies might actually be the natural course of epilepsy, not the placebo effect. Also, it is possible that if we understand this problem better, we may be able to better tell the difference between a “powerful” medicine, and a non-helpful one.

HOW THIS WILL HELP PATIENTS
If the studies can tell the difference between “natural” changes in seizure versus changes from “placebo effects” or medicines, then better medicines can be found faster and for less money.

HYPOTHESIS:
Review of clinical trials placebo arm patients will show evidence of natural seizure fluctuations rather than true placebo effects.

SPECIFIC AIMS:
1. To analyze clinical trial data from patients in reverse time. Instead of comparing baseline to treatment and looking for response, the comparison will be reversed. The expectation is placebo arm patients will continue to show “response” and active treatment arm patients will demonstrate “worsening”.
2. To pool clinical trial data from placebo arm patients in multiple placebo controlled epilepsy trials, and to clinically validate the findings in the recent simulation study.
3. To study the possibility of decreasing the “placebo effect” by retrospective review of the clinical data in a fashion analogous to that used in the recent simulation study.

PUBLICATION PLAN
The outcome of these investigations will be submitted for publication in a peer reviewed journal.

Study Data Provided

[{ "PostingID": 1803, "Title": "GSK-LAM100036", "Description": "A Multicenter, Double-blind, Randomized, Parallel-group Evaluation of LAMICTAL Extended-Release Adjunctive Therapy in Patients with Primary Generalized Tonic-Clonic Seizures

Medicine: lamotrigine, Condition: Epilepsy, Tonic-Clonic, Phase: 3, Clinical Study ID: LAM100036, Sponsor: GSK" },{ "PostingID": 1804, "Title": "GSK-LAM100034", "Description": "A Multicenter, Double-Blind, Randomized, Parallel-group Evaluation of LAMICTAL Extended-release Adjunctive Therapy in Subjects with Partial Seizures

Medicine: lamotrigine, Condition: Epilepsy, Partial, Phase: 3, Clinical Study ID: LAM100034, Sponsor: GSK" },{ "PostingID": 1814, "Title": "GSK-LAM40097", "Description": "A Multi-Center, Double-Blind, Randomized, Placebo-Controlled, Parallel-Group Evaluation of Lamotrigine Adjunctive Therapy in Subjects wtih Primary Generalized Tonic-Clonic Seizures

Medicine: lamotrigine, Condition: Epilepsy, Tonic-Clonic, Phase: 4, Clinical Study ID: LAM40097, Sponsor: GSK" },{ "PostingID": 2367, "Title": "UCB-SP667", "Description": "A Multicenter, Double-blind, Randomized, Placebo-controlled, Parallel
Group Trial to Investigate the Efficacy and Safety of SPM 927 (200mg/day, 400mg/day,
600mg/day) as Adjunctive Therapy in Subjects with Partial Seizures with or without
Secondary Generalization

Medicine: Lacosamide , Condition: Epilepsy, Phase: 2, Clinical Study ID: SP667, Sponsor: UCB" },{ "PostingID": 2368, "Title": "UCB-SP754", "Description": "SPM 927 (400mg/Day and 600mg/Day) as Adjunctive Therapy in Subjects With Partial Seizures With or Without Secondary Generalization

Medicine: Lacosamide , Condition: Epilepsy, Phase: 3, Clinical Study ID: SP754, Sponsor: UCB" },{ "PostingID": 2369, "Title": "UCB-SP755", "Description": "A Multicenter, Double-blind, Randomized, Placebo-controlled, Parallel Group Trial to Investigate the Efficacy and Safety of SPM 927 (200mg/Day and 400mg/Day) as Adjunctive Therapy in Subjects With Partial Seizures With or Without Secondary Generalization

Medicine: Lacosamide , Condition: Epilepsy, Phase: 3, Clinical Study ID: SP755, Sponsor: UCB" },{ "PostingID": 2370, "Title": "UCB-N01057", "Description": "Study to Evaluate the Safety and Efficacy of Levetiracetam in Patients Suffering From Idiopathic Generalized Epilepsy With Primary Generalized Tonic-clonic Seizures

Medicine: Levetiracetam , Condition: Epilepsy, Phase: 3, Clinical Study ID: N01057, Sponsor: UCB" },{ "PostingID": 2970, "Title": "GSK-VRX-RET-E22-301", "Description": "A Randomized, Double-Blind, Placebo-Controlled, Multicenter, Parallel-Group Phase 3 Study to Determine the Efficacy and Safety of Retigabine (1200mg/day) Used as Adjunctive Therapy in Refractory Epilepsy Patients with Partial-Onset Seizures

Medicine: retigabine, ezogabine, Condition: Seizures, Phase: 3, Clinical Study ID: VRX-RET-E22-301, Sponsor: GSK" },{ "PostingID": 2971, "Title": "GSK-VRX-RET-E22-302", "Description": "Randomized, Double-Blind, Placebo-Controlled, Multicenter, Parallel-Group Phase 3 Study - Determine Efficacy and Safety of Two Doses of Retigabine (900mg/Day and 600mg/Day) Used as Adjunctive Therapy in Refractory Epilepsy Patients with Partial-Onset Seizures

Medicine: retigabine, ezogabine, Condition: Seizures, Phase: 3, Clinical Study ID: VRX-RET-E22-302, Sponsor: GSK" },{ "PostingID": 2973, "Title": "GSK-3065A1-205", "Description": "A Randomized, Double-blind, Placebo-controlled, Parallel group, Multicenter, Dose-ranging, Efficacy and Safety Study of Retigabine (D-23129; GKE-841) Administered as Add-on Therapy in Patients With Partial Epilepsy

Medicine: retigabine, ezogabine, Condition: Seizures, Phase: 2/3, Clinical Study ID: 3065A1-205, Sponsor: GSK" }]

Statistical Analysis Plan

1. A random-effects meta-analysis of responder rates and/or risk ratios for placebo arm patients and active treatment patients when the time is reversed. In this analysis, individual study data will need to be re-calculated. To control for bias, we may adjust for the covariate of age as a continuous variable, or as a binary variable (age<18), or we will consider stratified meta-analysis for age. Comparing the placebo arm effect size estimate in forward time to the placebo arm effect size estimate in reverse time using a Fisher's Exact test, we can detect small differences that could be attributed to “true” placebo effects and regression to the mean. For example, if placebo effects and regression to the mean contribute 7% of the effect size, while natural fluctuation of disease contributes 8%, then forward size should be 15% and reverse size should be 8%. If we assume an average of 0.15 effect size (seen in multiple meta-analyses of placebo effects in epilepsy), and we wish to detect a drop of 5% in the effect size when time is reversed, the following power analysis applies. With 575 patients on treatment A (reverse time placebo) and 575 patients on treatment B (forward time placebo), there will be an 80% chance of detecting a significant difference at a one-sided 0.05 significance level. This assumes that the response rate of treatment A is 0.1 and the response rate of treatment B is 0.15. Therefore at least 575 placebo patients would be needed under these assumptions.
2. A pooled analysis of the responder rates for patients taken across studies (when possible) for simulated placebo trials that have baseline and test periods that are changed in multiple ways, as was performed in Goldenholz et al 2015.
3. An exploratory analysis of methods to decrease the placebo effect. Example methods to be explored will be delayed start, and cohort reduction. Other techniques may be explored as well. To assess the result, Fisher Exact test will be used at the individual study level, and meta-analysis will summarize the results across trials.

Patients with missing data will be excluded from the overall analysis.

Publication Citation

The publication citation will be added after the research is published.