Initial severity and antidepressant efficacy for anxiety disorders: an individual patient data meta-analysis

Proposal
1173

Title of Proposed Research

Lead Researcher

Ymkje Anna de Vries

Affiliation

Interdisciplinary Center Psychopathology and Emotion regulation, Department of Psychiatry, University Medical Center Groningen, the Netherlands

Funding Source

No specific funding will be used for the proposed project. All members of the research team are employed by and receive a salary from the University Medical Center Groningen.

Potential Conflicts of Interest

Potential conflicts of interest will be disclosed when the research is presented and published.

Data Sharing Agreement Date

30 June 2015

Lay Summary

Anxiety disorders form a heterogeneous group of disorders characterized by excessive fear or worry. The group includes generalized anxiety disorder (GAD), social anxiety disorder (SAD), obsessive-compulsive disorder (OCD), post-traumatic stress disorder (PTSD) and panic disorder (PD). In any given year, 18.1% of the US population will suffer from an anxiety disorder, while the lifetime risk stands at 28.8%. As these disorders are also characterized by an early age of onset and a chronic course, they pose a major public health burden. Anxiety disorders are often treated with medications. In particular, second-generation antidepressants, which include selective serotonin reuptake inhibitors (SSRIs) such as fluoxetine (Prozac) and serotonin-norepinephrine reuptake inhibitors (SNRIs), have been found to be effective treatments and are now considered the pharmacological treatment of first choice for anxiety.However, much of what is known about antidepressants is derived from research in depression rather than anxiety. In recent years, researchers have found that the efficacy of antidepressants for depression depends upon the initial severity of symptoms. While severely depressed patients experience considerable benefit from taking antidepressants, as compared to placebo, mildly depressed patients experience a much smaller benefit. Consequently, it has been recommended that antidepressants should not be a first-line treatment for mild depression.It is possible that a similar relationship between symptom severity and antidepressant efficacy exists for anxiety disorders. As antidepressants are frequently prescribed for patients with mild or moderate anxiety, a clear understanding of their effectiveness across the severity range is vital to inform treatment decisions. The evidence, however, is limited. We have previously conducted a meta-analysis of all premarketing studies of second-generation antidepressants that are approved in the US for the short-term treatment of anxiety. In this meta-analysis, we found no evidence that initial severity affects antidepressant efficacy for anxiety disorders, in contrast to findings in depression. However, as we relied upon trial-level rather than patient-level data, definitive conclusions cannot be drawn yet. In the proposed project, we therefore plan to use individual patient data to examine whether initial symptom severity affects antidepressant efficacy for anxiety disorders. We will make use of data from 30 randomized controlled trials of antidepressants (specifically: duloxetine, fluoxetine, and paroxetine) for GAD, SAD, OCD, PTSD or PD. We will model improvement in drug and placebo groups over time and investigate whether greater initial severity predicts a greater difference in improvement between drug-treated and placebo-treated patients. Our findings will be presented at international conferences and published in a major medical journal.

Study Data Provided

[{ "PostingID": 1647, "Title": "LILLY-F1J-MC-HMBR", "Description": "Duloxetine Hydrochloride 60 mg or 120 mg Once Daily Compared With Placebo in Patients With Generalized Anxiety Disorder" },{ "PostingID": 2465, "Title": "LILLY-F1J-MC-HMDT", "Description": "Duloxetine Hydrochloride Once Daily Compared With Placebo in the Treatment of Generalized Anxiety Disorder" },{ "PostingID": 2466, "Title": "LILLY-F1J-MC-HMDU", "Description": "A Comparison of Duloxetine Hydrochloride, Venlafaxine Extended Release, and Placebo in the Treatment of Generalized Anxiety Disorder." },{ "PostingID": 2467, "Title": "LILLY-F1J-MC-HMDW", "Description": "A Comparison of Duloxetine Hydrochloride, Venlafaxine Extended Release, and Placebo in the Treatment of Generalized Anxiety Disorder" },{ "PostingID": 2469, "Title": "LILLY-B1Y-EW-HCHQ", "Description": "A Double-Blind Placebo-Controlled Study of Fluoxetine and Clomipramine in the Treatment of Panic Disorder" },{ "PostingID": 2470, "Title": "LILLY-B1Y-MC-HCEP", "Description": "Fluoxetine vs Placebo in Obsessive Compulsive Disorder" },{ "PostingID": 2471, "Title": "LILLY-B1Y-MC-HCHG", "Description": "Fluoxetine versus Placebo in Panic Disorder" },{ "PostingID": 2472, "Title": "LILLY-B1Y-MC-HCJB", "Description": "Fluoxetine versus Placebo in Panic Disorder" },{ "PostingID": 2473, "Title": "LILLY-B1Y-MC-HCJC", "Description": "Fluoxetine versus Placebo in Panic Disorder" },{ "PostingID": 2523, "Title": "GSK-BRL-029060/CPMS- 116", "Description": "Paroxetine versus Placebo in the Treatment of Obsessive-Compulsive Disorder" },{ "PostingID": 2524, "Title": "GSK-MY- 1028/BRL-029060/1/CPMS-118", "Description": "Paroxetine versus Clomipramine and Placebo in the Treatment of Obsessive-Compulsive Disorder" },{ "PostingID": 2525, "Title": "GSK-MY1047/BRL-029060/1/CPMS-120", "Description": "A double-blind, placebo-controlled, multicenter study of fixed doses of paroxetine (10, 20, and 40mg) given as a single oral dose daily, in the treatment of panic disorder." },{ "PostingID": 2526, "Title": "GSK-MY-1037/BRL-029060/1/CPMS-136", "Description": "A Double-Blind Study to Assess the Efficacy and Tolerance of a Flexible Dose of Paroxetine Compared with a Flexible Dose of Clomipramine and Placebo in the Treatment of Obsessive Compulsive Disorder." },{ "PostingID": 2527, "Title": "GSK-MY-1036/BRL-029060/1/CPMS-187 (PAR187)", "Description": "A Double-Blind Placebo Controlled Comparative Study of Paroxetine and Clomipramine in the Treatment of Panic Disorder." },{ "PostingID": 2528, "Title": "GSK-MY-1048/BRL-029060/1/CPMS-223 (PAR 223)", "Description": "A Double-Blind, Multicentered, Flexible-Dose Study of Paroxetine, Alprazolam and Placebo in the Treatment of Panic Disorder" },{ "PostingID": 2529, "Title": "GSK-PAR 029060-382", "Description": "A Randomized, Double-Blind, Comparison of Paroxetine and Placebo in the Treatment of Generalized Social Phobia" },{ "PostingID": 2530, "Title": "GSK-PAR 454", "Description": "A Randomized, Double-Blind, Fixed Dose Comparison of 20, 40, and 60mg Daily of Paroxetine and Placebo in the Treatment of Generalized Social Phobia." },{ "PostingID": 2531, "Title": "GSK-29060-494", "Description": "A Double-Blind, Placebo-Controlled, Flexible-Dosing Trial to Evaluate the Efficacy of Modified-Release Paroxetine in the Treatment of Panic Disorder" },{ "PostingID": 2532, "Title": "GSK-29060/495", "Description": "A Double-Blind, Placebo-Controlled, Flexible-Dosing Trial to Evaluate the Efficacy of Controlled-Release Paroxetine in the Treatment of Panic Disorder" },{ "PostingID": 2533, "Title": "GSK-29060/497", "Description": "A Double-Blind, Placebo-Controlled, Flexible-Dosing Trial to Evaluate the Efficacy of Controlled-Release Paroxetine in the Treatment of Panic Disorder" },{ "PostingID": 2534, "Title": "GSK-PAR 029060-502", "Description": "A Randomised, Double-blind Study of Paroxetine and Placebo in the Treatment of Social Phobia" },{ "PostingID": 2535, "Title": "GSK-PAR 29060/627", "Description": "A 12-Week, Double-Blind, Placebo-Controlled, Parallel Group Study to Assess the Efficacy and Tolerability of Paroxetine in Patients Suffering from Post-traumatic Stress Disorder (PTSD)" },{ "PostingID": 2536, "Title": "GSK-PAR 29060 637", "Description": "A double-blind, placebo controlled study to evaluate the efficacy and tolerability of paroxetine in patients with Generalised Anxiety Disorder (GAD)" },{ "PostingID": 2537, "Title": "GSK-BRL-029060/RSD-101336/1/CPMS-641", "Description": "A Randomized, Double-Blind, Placebo Controlled, Fixed Dosage Trial to Evaluate the Efficacy and Tolerability of 20 and 40mg/day Paroxetine in Patients with Generalized Anxiety Disorder" },{ "PostingID": 2538, "Title": "GSK-PAR 29060/642", "Description": "A Randomized, Double-Blind, Placebo Controlled, Flexible Dosage Trial to Evaluate the Efficacy and Tolerability of Paroxetine in Patients with Generalized Anxiety Disorder" },{ "PostingID": 2539, "Title": "GSK-PAR 648", "Description": "A 12 Week, Double-blind, Placebo-controlled, parallel group study to assess the efficacy and tolerability of Paroxetine in Patients suffering from Post-traumatic Stress Disorder (PTSD)" },{ "PostingID": 2540, "Title": "GSK-PAR 29060 651 (BRL – 02960)", "Description": "A 12 week, double-blind, fixed dose comparison of 20 and 40mg daily of paroxetine and placebo in Patients suffering from Post-traumatic Stress Disorder (PTSD)." },{ "PostingID": 2541, "Title": "GSK-BRL-029060/CPMS-790", "Description": "A Double-blind, Placebo-controlled, Flexible-dose Study of Paroxetine CR in the Treatment of Patients with Social Anxiety Disorder" },{ "PostingID": 2542, "Title": "GSK-29060/791", "Description": "A Randomized, Double-Blind, Placebo-Controlled, Flexible Dosage Trial to Evaluate the Efficacy and Tolerability of Paroxetine CR in Patients with Generalized Anxiety Disorder (GAD)" }]

Statistical Analysis Plan

A separate longitudinal analysis will be conducted for each disorder. For GAD, SAD, OCD, and PTSD, we will use linear mixed models; for PD, we will use a generalized linear mixed model (multilevel negative binomial regression), as the dependent variable for this disorder (number of panic attacks) is a discrete count variable. We will use maximum likelihood estimation as the estimation method for the linear mixed models, while Laplace approximation will be used as the estimation method for the multilevel negative binomial regression for PD. In all models, measurement occasion represents level 1, participants represent level 2, and trial represents level 3. The effect measure of interest is the change in symptoms from baseline, except for PD, for which the effect measure of interest is the total number of panic attacks per two weeks.

We will build an initial model by including all the fixed effects of interest, regardless of significance. These include initial severity, treatment group and covariates (see below). Linear and quadratic terms for time (in days since baseline) will be included. For each participant, the actual visit dates will be used (if available) rather than the intended weekly visit date. We will also include the following two- and three-way interactions: severity x group, severity x linear time, group x linear time, severity x group x linear time, severity x quadratic time, group x quadratic time, severity x group x quadratic time.

Using this first model, we will model the variance-covariance structure of the nested data by including random effects. Random effects for (linear and quadratic) time, group, severity and interactions will be considered, as well as autocorrelated errors. Restricted maximum likelihood (REML) will be used for estimation, and the best-fitting model will be selected based upon the Akaike Information Criterion (AICc).

In this best-fitting model, we will use backward selection with maximum likelihood (ML) to select the significant fixed effects. Non-significant interaction terms will be removed from the model (unless the three-way interaction of group x severity x (linear or quadratic) time is significant, in which case all two-way interactions and main effects that use these variables will be retained). The best-fitting model will again be selected based upon the Akaike Information Criterion (AICc).

Covariates
We will test models with and without the following covariates: age, gender, and duration of illness. Only the main effect of these covariates will be included; no interactions with other variables will be included.

Power
For each disorder, we have at least 1,100 participants available (with a minimum of 350 participants per group). Furthermore, as each trial has 5 to 8 post-baseline measurement occasions, the total amount of data points may be as much as 5,000 or more per disorder (depending on the dropout rate). Given the large amount of data, power to detect a meaningful effect will be sufficient.

Missing data
We expect baseline variables (such as initial severity) to be essentially complete, but some change scores are likely to be missing due to dropout or missed measurement occasions. We assume that these data are missing at random (MAR), that is: missingness of the dependent variable may depend upon observed variables (such as previous symptom scores or covariates), but it does not depend upon the value of the unobserved (missing) variable. When MAR holds, the mixed model yields unbiased estimates of coefficients and standard errors even when some data is missing, and no other methods for handling missing data are required.

Publication Citation

https://doi.org/10.1002/da.22737