Untitled

A pooled analysis of two placebo-controlled trials ofdesvenlafaxine in major depressive disorderDaniel Z. Liebermana, Stuart A. Montgomeryb, Karen A. Tourianc,Claudine Brisardd, Gregory Rosasc, Krishna Padmanabhanc,Jean-Michel Germaind and Bruno Pitroskyd The efficacy, safety, and tolerability of desvenlafaxine and placebo, respectively; magnitude of effect = – 0.37 (administered as desvenlafaxine succinate) were evaluated (P < 0.001)]. Adverse events were comparable to those in two similarly designed, phase 3, randomized, double- found with other drugs sharing a similar mechanism of blind, placebo-controlled, venlafaxine-extended-release- action. These data support the efficacy, safety, and referenced, flexible-dose studies of outpatients with a tolerability of desvenlafaxine in the treatment of major primary diagnosis of major depressive disorder. Owing to a depressive disorder. Int Clin Psychopharmacol 23:188–197 high placebo response, the individual studies were 2008 Wolters Kluwer Health | Lippincott Williams & underpowered. Therefore, a post-hoc pooled analysis was performed (desvenlafaxine and placebo data were pooled; venlafaxine extended release data were not, owing to International Clinical Psychopharmacology 2008, 23:188–197 different flexible-dose regimens in the two studies). The Keywords: desvenlafaxine, major depressive disorder, mixed-effect model primary outcome measure was the change from baseline on the 17-item Hamilton Rating Scale for Depression; the Clinical Global Impressions-Improvement item score was a aGeorge Washington University Clinical Psychiatric Research Center,Washington, DC, bImperial College School of Medicine, London, UK, secondary outcome. Analysis of the pooled data (using a cWyeth Research, Collegeville, Pennsylvania, USA and dWyeth Research, mixed-effect model for repeated measures) revealed that after 8 weeks of treatment, desvenlafaxine was Correspondence to Dr Daniel Z. Lieberman, MD, George Washington University significantly better than placebo on 17-item Hamilton Clinical Psychiatric Research Center, 2150 Pennsylvania Avenue, NW, Rating Scale for Depression [ – 14.21 vs. – 11.87 for desvenlafaxine and placebo, respectively; magnitude of Tel: + 1 202 741 2899; fax: + 1 202 741 2891; e-mail: cfrdzl@gwumc.edu effect = – 2.34 (P < 0.001)] and Clinical Global Impressions- Received 31 October 2007 Accepted 27 March 2008 Improvement item scores [1.95 vs. 2.32 for desvenlafaxine because preclinical studies have demonstrated that it Depression, a common mental disorder, is a major cause inhibits the neuronal uptake of both serotonin and of disability throughout the world, and a serious public norepinephrine and, to a lesser degree, dopamine health concern. Worldwide, more than 150 million people reuptake (Muth et al., 1991; Clement et al., 1998). It suffer from depression, and nearly 1 million commit does not have any monoamine oxidase inhibitory activity, suicide every year (World Health Organization, 2001).
and shows virtually no affinity for rat brain muscarinic, Even with treatment, a large percentage of patients, who cholinergic, H1-histaminergic, or a1-adrenergic receptors receive currently available therapies, recover only par- tially, often with continued functional impairment owingto residual symptoms, underscoring the importance ofand need for novel antidepressants (Steffens et al., 1997; Desvenlafaxine has been examined in a series of Thase et al., 2001; Faravelli et al., 2003; Segal et al., 2003; preclinical in-vivo and in-vitro tests and has been found to be active in multiple models used to predictantidepressant activity (Alfinito et al., 2006). Results of Desvenlafaxine (administered as desvenlafaxine succi- two phase 3 clinical trials showed that desvenlafaxine nate) is the major active metabolite of venlafaxine.
had significantly better efficacy compared with placebo Desvenlafaxine is an antidepressant that has been based on the 17-item Hamilton Rating Scale for Depres- approved by the US Food and Drug Administration for the treatment of major depressive disorder (MDD).
Global Impressions-Improvement Scale (CGI-I) (Guy, Desvenlafaxine is chemically unrelated to tricyclic, tetracyclic, or other available antidepressants (with the exception of venlafaxine) and is classified as a dual-acting serotonin and norepinephrine reuptake inhibitor (SNRI) c 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Desvenlafaxine in the treatment of major depressive disorder Lieberman et al.
The two phase 3 studies discussed herein, which com- pared the antidepressant efficacy of desvenlafaxine Following the screening period (6–14 days), eligible and placebo in MDD, are the only studies that included patients were randomly assigned to one of the treatment venlafaxine extended release (ER) as a reference treat- groups (desvenlafaxine, venlafaxine ER, or placebo) and ment. In retrospect, these studies were underpowered, received up to 8 weeks of treatment. Patients randomized largely owing to the high placebo response rate observed to desvenlafaxine were treated with an initial target dose in both the studies. Therefore, to examine the efficacy of 200 mg/day. Patients in the EU study were started of desvenlafaxine in an adequately powered analysis, on the initial target dose on day 1, whereas those in the the data from both of these studies were pooled post hoc.
US study received 4 days of 100 mg/day before reaching The results of the pooled analysis and the primary results the target dose. Each study had an optional increase to of the individual studies are presented in this article.
400 mg/day after day 28 or decrease back to 200 mg/dayat any time, based on the investigator’s judgment. At theend of the study, the patients underwent a taper period Two similar studies were performed, one in Europe (EU)and one in the United States (US). Each was a double- Venlafaxine ER was used as an active control. In the EU blind, multisite, placebo-controlled, parallel-group, ven- study, patients assigned to venlafaxine ER received lafaxine ER-referenced, flexible-dose trial designed to 75 mg/day for 28 days, with an optional increase to compare the antidepressant efficacy, safety, and toler- 150 mg/day after day 28, based on the investigator’s ability of desvenlafaxine with placebo. The use of placebo judgment. In the US study, patients received a daily dose was necessary to provide reliable scientific evidence of of 75 mg/day for 4 days. This dose was increased to 150 mg/day on day 5, and there was an optional increaseto 225 mg/day after day 28. At the end of the study The studies were approved by an independent ethics period, the patients underwent a taper period based on committee or Institutional Review Board and were consistent with Principles of Good Clinical Practice andapplicable regulatory requirements in each participating Efficacy, safety, and tolerability evaluations country. All participants provided written informed The primary efficacy measure was the HAM-D17 total score, ascertained at each visit. Secondary efficacymeasures included the CGI-I score, the response rate as measured by a 50% or greater decrease in the score on Men and women, outpatients 18–75 years of age with a the HAM-D17, the percentage of patients in remission primary diagnosis of MDD, based on a psychiatric inter- (HAM-D17 scores of 7 or less), MADRS total score, view using the ‘Diagnostic and Statistical Manual of CGI-S score, the Visual Analog Scale-Pain Intensity Mental Disorders, Fourth Edition’ (American Psychiatric (VAS-PI) (DeLoach et al., 1998) overall pain and subscale Association, 1994) criteria, single or recurrent episode, scores, HAM-D6 [Bech version (Bech et al., 1975)] total without psychotic features, were eligible for study score, Covi Anxiety Scale total score, and response rates participation. At baseline and screening, patients were also required to have a minimum HAM-D17 score of22 and score at least 2 on item one (depressed mood) of Safety and tolerability were determined using the HAM-D17, a Clinical Global Impressions-Severity (CGI-S) following assessments: monitoring of adverse events Scale (Guy, 1976) score of at least 4, and a Raskin (AEs), discontinuation because of AEs, physical examina- Depression Scale (Raskin et al., 1969) score greater than tion, standard 12-lead electrocardiogram, vital signs the Covi Anxiety Scale (Lipman, 1982) score.
(weight, pulse, and blood pressure), and laboratory deter-minations (hematology, blood chemistry, and urinalysis).
The screening evaluation included a medical history anda psychiatric history. The modified Mini International Neuropsychiatric Interview was used as a secondary The study designs of the US and EU studies were similar confirmation of the primary diagnosis of MDD and any with respect to duration of treatment, desvenlafaxine comorbid psychiatric disorders that may have been daily dose, and efficacy measures; the dosing schedules present. Patients with comorbid substance use disorders of venlafaxine ER were different. In both the studies, were excluded; however, patients with comorbid general- desvenlafaxine was dosed from 200 to 400 mg/day, ized anxiety disorder, panic disorder, or social anxiety whereas the venlafaxine ER dosing was 75–150 mg/day disorder were allowed to participate as long as MDD was in the EU study and 150–225 mg/day in the US study.
the primary diagnosis. Patients at high risk for suicidal The flexible dosing schedule of desvenlafaxine and placebo data allowed for pooling. However, the venlafax- Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
International Clinical Psychopharmacology ine ER dosing schedule was lower in the EU study (75–150 mg/day) than in the US study (150–225 mg/day).
Therefore, the venlafaxine ER data were not pooled and A total of 738 patients were randomly assigned to are presented as two groups, allowing for a four-arm, treatment (250 to placebo, 239 to desvenlafaxine, 128 to pooled analysis. As each study was designed to compare venlafaxine ER 75–150 mg/day, and 121 to venlafaxine desvenlafaxine with placebo, the differences in the ER 150–225 mg/day). Eighteen patients had no data after dosing of the active control did not affect the data baseline and were not included in the safety population.
analysis of the efficacy of desvenlafaxine.
The remaining 720 patients, who completed the prestudyperiod and took the randomly assigned study drug underdouble-blind conditions, were included in the safety All efficacy analyses were based on the intent-to-treat analyses. Seven patients of the safety population did (ITT) population, or efficacy population, which consisted not meet criteria for the ITT population, which included of all randomized patients who had a baseline primary 713 patients (226 desvenlafaxine, 127 venlafaxine ER evaluation, who took at least one dose of study drug, and 75–150 mg/day, 115 venlafaxine ER 150–225 mg/day, and who had at least one primary efficacy evaluation after 245 placebo). There were 574 completers (166 desvenla- the first dose of the study drug. Safety and tolerability faxine, 108 venlafaxine ER 75–150 mg/day, 90 venlafaxine analyses were based on the safety and tolerability ER 150–225 mg/day, and 210 placebo).
population, which included all randomized patients whohad taken at least one dose of the study drug.
Demographic and other baseline characteristicsDemographic and baseline clinical characteristics of the pooled ITT population were generally similar (Table 1).
Longitudinal changes from baseline on pooled data for Reflecting the geographic differences between the two pooled studies, minor differences in demographic char- secondary measures (CGI-I, CGI-S, MADRS, HAM-D acteristics were seen among the patients. Individuals in total, Covi Anxiety Scale, and VAS-PI) were analyzed the EU study had a lower mean weight than those in the using a mixed-effect model for repeated measures US study, and a greater percentage of patients were non- (MMRM) analysis. For the MMRM analysis, the change Hispanic whites. As the magnitude of these differences from baseline on the respective scale (except for CGI-I) was small, they did not interfere with the use of the was analyzed as the outcome variable. An autoregressive pooled analysis. Mean baseline severity on the HAM-D17 first order [AR(1)] correlation structure was used to ranged from 25.1 to 25.8 and did not show statistically model the within correlation, with treatment groups, significant differences between groups.
weeks, and the treatment-group-by-week interaction asfixed factors, center as a random factor, and baseline as a covariate. For the CGI-I, the score on the CGI-I Scale was used as the outcome. Changes from baseline to end point for secondary outcomes were also analyzed with an In the pooled analyses using MMRM, a significant analysis of covariance (ANCOVA) with terms of treat- difference in HAM-D17 total scores between desvenla- ment and study (protocol) as factors and the baseline faxine and placebo was observed at week 3 and was score as a covariate; last-observation-carried-forward maintained throughout the treatment period (Fig. 1, (LOCF) and observed cases (OC) data were analyzed.
Table 2). By the week 8 evaluation, the change from A logistic regression model was used for binary outcome baseline in the HAM-D17 total score was – 14.21 for variables (response and remission measured by the desvenlafaxine versus – 11.87 for placebo (magnitude of effect = – 2.34; P < 0.001) (Fig. 2a). Both venlafaxine ER 17 and CGI-I). Treatment effects were tested at a two-sided significance level of 0.05.
groups were significantly different from placebo at theweek 8 evaluation (75–150 mg/day: – 14.26, P = 0.001;150–225 mg/day: – 14.56, P < 0.001); the venlafaxine ER 75–150 mg/day group showed significant separation from In the individual studies the primary efficacy measure, placebo beginning at week 6, whereas the 150–225 mg/ HAM-D17 total score, was evaluated using ANCOVA on day dose of venlafaxine ER showed more rapid efficacy changes from baseline with the treatment arm and site as the factors and baseline scores as the covariate. LOCF(primary analysis) and OC analyses were both performed.
Statistically significant differences in CGI-I scores between The CGI-I score was analyzed by using analysis of variance desvenlafaxine and placebo were observed at week 8 (ANOVA) with the treatment arm and site as the factors.
(2.0 vs. 2.3; P < 0.001; MMRM analysis) (Fig. 2b); Response and remission rates on the HAM-D17 were significant separation from placebo was also observed at analyzed with the logistic regression model, with treatment weeks 3 (2.6 vs. 2.8; P = 0.014), 4 (2.3 vs. 2.7; P < 0.001), and site as the factors and baseline score as a covariate.
and 6 (2.1 vs. 2.4; P < 0.001). Both doses of venlafaxine Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Desvenlafaxine in the treatment of major depressive disorder Lieberman et al.
Demographic and baseline characteristics: ITT pooled population CGI-S, Clinical Global Impressions-Severity; DVS, desvenlafaxine; HAM-D17, 17-item Hamilton Rating Scale for Depression; ITT, intent to treat; VEN ER, venlafaxineextended release.
Change in HAM-D17 total score over time (MMRM), ITT pooled and individual populations (desvenlafaxine and placebo data represent pooled data from two studies; venlafaxine ER data were not pooled owing to differences in dosing). Mean doses: study 309: 302 mg desvenlafaxine; 118 mg venlafaxine ER Study 317: 336 mg desvenlafaxine; 206 mg venlafaxine ER. aP < 0,01 venlafaxine ER 150– 225 mg versus placebo; bP < 0,05 desvenlafaxine versus placebo; cP < 0.001 desvenlafaxine versus placebo; dP < 0,01 venlafaxine ER 75–150 mg versus placebo; eP < 0.001 venlafaxine ER 150–225 mg versus placebo. ER, extended release; HAM-D Rating Scale for Depression; MMRM, mixed-effect model for repeated ER also were significantly different from placebo at week 8(75–150 mg/day: 2.0, P = 0.003; 150–225 mg/day: 1.9, DVS, desvenlafaxine; HAM-D17, 17-item Hamilton Rating Scale for Depression;ITT, intent to treat; MMRM, mixed-effect model for repeated measures; SE, P < 0.001) (Fig. 2b). In the ANCOVA analysis, statistically standard error; VEN ER, venlafaxine extended release.
significant differences were observed with desvenlafaxine aDesvenlafaxine and placebo data represent pooled data from two studies; venlafaxine ER data were not pooled owing to differences in dosing.
and for venlafaxine 150–225 mg/day on both the LOCF(final evaluation) and week 8 OC analyses of these out-comes (data not shown).
response rates were observed for both venlafaxine ERdose groups (75–150 mg/day: 64%, P = 0.033; 150– At the final evaluation, 55% of desvenlafaxine patients 225 mg/day: 57%, P = 0.017) and in remission rates for were HAM-D17 responders ( Z 50% reduction in HAM-D17 the 150–225 mg/day dose group (75–150 mg/day: 34%, total score) compared with 47% of patients who had P = 0.133; 150–225 mg/day: 36%, P = 0.003) (Fig. 3).
received placebo. Thirty percent of patients on desven- CGI-I response rates were not significantly different for lafaxine achieved remission (HAM-D17 total score r 7) desvenlafaxine (62%) compared with placebo (56%); compared with 23% of patients on placebo. The response rates were statistically significant for the differences in HAM-D response and remission rates for venlafaxine ER 75–150 mg/day group (73%; P = 0.039) desvenlafaxine versus placebo were not statistically and 150–225 mg/day group (65%; P = 0.011) compared significant; significant differences from placebo in Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
International Clinical Psychopharmacology HAM-D17 response and remission rates, final evaluation: pooled andindividual populations (desvenlafaxine and placebo data representpooled data from two studies; venlafaxine ER data were not pooled owing to differences in dosing). Response Z 50% reduction in HAM-D17 total score. Remission = HAM-D17 total score r 7. ER, extendedrelease; HAM-D17, 17-item Hamilton Rating Scale for Depression.
(P < 0.001), and arm, leg, or joint pain (P = 0.002).
Results of the ANCOVA analysis (LOCF final evaluation) reflected statistically significant differences from placebo on fewer outcomes, for all treatment groups (desvenla- faxine: HAM-D6 and VAS-PI back pain; venlafaxine75–150 mg/day: VAS-PI overall pain, back pain, and arm, leg, and joint pain; venlafaxine 150–225 mg/day: MADRS, (a) Efficacy results, MMRM, week 8 evaluation, HAM-D17: pooled and CGI-S, HAM-D6, and VAS-PI arm, leg, and joint pain); individual populations (desvenlafaxine and placebo data represent analysis of week 8 OC data showed the following pooled data from two studies: venlafaxine ER data were not pooled statistically significant differences from placebo: desven- owing to differences in dosing). aP < 0.01. (b) Efficacy results, MMRM,week 8 evaluation, CGI-I: pooled and individual populations lafaxine (MADRS, CGI-S, HAM-D6, Covi Anxiety, VAS- (desvenlafaxine and placebo data represent pooled data from two PI overall pain, back pain, and chest pain), venlafaxine studies; venlafaxine ER data were not pooled owing to differences indosing). aP < 0.01. CGI-I, Clinical Global Impressions-Improvement; 75–150 mg/day (VAS-PI overall pain and arm, leg, and ER, extended release; HAM-D17, 17-item Hamilton Rating Scale for joint pain), and venlafaxine 150–225 mg/day (MADRS, Depression; MMRM, mixed-effect model for repeated measures.
CGI-S, HAM-D6, VAS-PI overall pain, back pain, andarm, leg, and joint pain).
Week 8 MMRM data for secondary efficacy measures are presented in Table 3. Desvenlafaxine was significantly Individual study results are presented in Tables 4 and 5.
superior to placebo on the majority of the secondary Owing to a high placebo effect, these individual studies outcomes: MADRS (P < 0.001), CGI-S (P < 0.001), were underpowered to show an effect of desvenlafaxine.
HAM-D6 (P < 0.001), Covi Anxiety Scale (P = 0.035), As such, desvenlafaxine did not significantly separate VAS-PI overall pain (P = 0.003) and the back pain from placebo on the primary analysis of either the (P < 0.001), chest pain (P = 0.037), and arm, leg, or HAM-D17 total score or CGI-I score (Table 4). Addition- joint pain (P = 0.027) subscale scores. Venlafaxine ER ally, HAM-D17 response and remission rates were not 75–150 mg/day was superior to placebo on symptom significantly different for desvenlafaxine compared with improvement as measured by the MADRS (P < 0.001), placebo (Table 5). There were statistically significant CGI-S (P < 0.001), HAM-D6 (P = 0.002), VAS-PI overall differences in HAM-D17 and CGI-I scores between the pain (P < 0.001) and back pain (P = 0.029), chest pain venlafaxine and placebo groups in the US study only (P = 0.023), and arm, leg, or joint pain (P < 0.001) (P = 0.005 and P = 0.011, respectively). Response rates subscales. Venlafaxine ER 150–225 mg/day was better with venlafaxine ER were statistically significant in both studies (P = 0.03 and P = 0.038 in EU and US, respec- tively); remission rates were significant only in the US (P = 0.015), VAS-PI overall pain (P = 0.004), back pain Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Desvenlafaxine in the treatment of major depressive disorder Lieberman et al.
Secondary efficacy endpoints, week 8 (MMRM analysis); The most common treatment-emergent adverse events included nausea, somnolence, dry mouth, and sweating (Table 6). The type and frequency of treatment- emergent adverse events reported were similar to those reported with other SNRIs. In the desvenlafaxine group, there were increases in mean serum lipids, blood pressure, and pulse compared with placebo and decreases in mean weight. No deaths occurred in either study.
The results of this analysis confirm and extend the results of the earlier phase 3 studies of desvenlafaxine (DeMartinis et al., 2007; Liebowitz et al., 2007; Septien-Velez et al., 2007).
The first of these studies (n = 461) showed a significant reduction in the HAM-D17 scores for the desvenlafaxine 100-mg (P = 0.0038) and 400-mg (P = 0.0023) dose groups versus the placebo group, and a trend toward significance (P = 0.076) in the 200-mg dose group. All desvenlafaxine dose groups showed significant improvement on the CGI-I Scale, a secondary efficacy measure, compared with placebo (P < 0.05) (DeMartinis et al., 2007). In the second phase 3 trial (n = 369), the adjusted mean change from baseline in the HAM-D17 total score, the primary efficacy measure, was significantly greater for the desvenlafaxine 200-mg (P = 0.002) and 400-mg (P = 0.008) dose groups versus placebo (Septien-Velez et al., 2007). A third, recently published study of desvenlafaxine (n = 234) that used a flexible-dose regimen – treatment was initiated at 100 mg/day for 14 days, after which the dose was increased to 200 mg/day, with an option to decrease the dose only if necessary for safety or tolerability – did not demonstrate a statistically significant difference between desvenlafaxine and placebo on the primary efficacy measure (difference in adjusted means = 1.0; P = 0.277), although significant differences were observed on some secondary measures (i.e. MADRS and VAS-PI overall pain, back pain, and arm, leg, or joint pain scales) (Liebowitz et al., 2007).
CGI-S, Clinical Global Impressions-Severity; CI, confidence interval; Covi, CoviAnxiety Scale; DVS, desvenlafaxine; HAM-D6, 6-item Bech version of 17-item Pooling of data from inconclusive placebo-controlled Hamilton Rating Scale for Depression; MADRS, Montgomery A˚sberg DepressionRating Scale; MMRM, mixed-effect model for repeated measures; VAS-PI, Visual studies provides a useful method of establishing whether Analog Scale-Pain Intensity; VEN ER, venlafaxine extended release.
the treatment effect observed was significant, provided aDesvenlafaxine and placebo data represent pooled data from two studies; the studies are comparable. In this case, the design of the venlafaxine ER data were not pooled owing to differences in dosing.
studies was similar, using the same duration and the sameprimary and secondary efficacy measures. The flexibledosage regime for desvenlafaxine was the same with the Although the EU study failed on the primary efficacy end target treatment dose dependent on efficacy and toler- point using ANCOVA (LOCF), significant differences ability, and this allowed for pooling of data on desvenla- were observed when alternative analytic methods were faxine and placebo. By contrast, the dosage regime for used; desvenlafaxine and venlafaxine ER were signifi- venlafaxine ER was different in each study, with a low cantly better than placebo at week 8 for ANCOVA (OC daily dose of 75–150 mg in the EU study and a higher data; P < 0.001 and P = 0.027, respectively) and MMRM daily dose of 150–225 mg in the US study; data from (P < 0.001 and P = 0.005, respectively). In the US study, these treatment arms were not combined. The pooled only venlafaxine ER separated from placebo, using both analysis therefore allowed for valid conclusions for the ANCOVA (LOCF and week 8 OC) and MMRM.
efficacy of desvenlafaxine, but not for venlafaxine ER as a Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
International Clinical Psychopharmacology Efficacy results (LOCF), final evaluation: ITT individual studies ANCOVA, analysis of covariance; ANOVA, analysis of variance; CGI-I, Clinical Global Impressions-Improvement; CI, confidence interval; DVS, desvenlafaxine; EU,Europe; HAM-D17, 17-item Hamilton Rating Scale for Depression; ITT, intent to treat; LOCF, last observation carried forward; US, United States; VEN ER, venlafaxineextended release.
HAM-D17 response and remission rates, final evaluation: individual studies CI, confidence interval; DVS, desvenlafaxine; EU, Europe; HAM-D17, 17-item Hamilton Rating Scale for Depression; US, United States; VEN ER, venlafaxine extendedrelease.
Most common TEAEs ( Z 5% and at least two times greater with DVS than with placebo): pooled and individual populationsa, DVS, desvenlafaxine; TEAEs, treatment-emergent adverse events; VEN ER, venlafaxine extended release.
aDesvenlafaxine and placebo data represent pooled data from two studies; venlafaxine ER data were not pooled owing to differences in dosing.
bIncidence based on the number of men: placebo = 85, desvenlafaxine = 75, and venlafaxine ER = 71.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Desvenlafaxine in the treatment of major depressive disorder Lieberman et al.
whole. In the pooled analysis of the two underpowered might not be the best in all cases (Mallinckrodt et al., studies described in this report, desvenlafaxine was 2004). The MMRM analysis is a type of likelihood-based, effective compared with placebo on both primary and mixed-effects method in which missing points are secondary efficacy measures. In particular, the efficacy estimated based on observed data. MMRM approaches of desvenlafaxine was reflected in the significant change are easy to implement, are more robust to the biases from in HAM-D17 and CGI-I scores compared with placebo. In missing data, and provide better control of type I and addition, desvenlafaxine was associated with a significant type II errors than LOCF ANOVA (Mallinckrodt et al., improvement compared with placebo on the HAM-D6 2004; Molenberghs et al., 2004). The differences in subscale, which assesses the core symptoms of depres- results between analytic methods a clearly demonstrated sion, the MADRS, CGI-S, Covi Anxiety Scale, and three by the lack of statistical separation in the EU and US of four VAS-PI subscales as well as overall pain score.
studies, where in the pooled analyses, using MMRM,desvenlafaxine was significantly more effective thanplacebo as measured by the HAM-D The individual studies and pooled analysis were not designed or powered to directly compare desvenlafaxine univariate repeated-measures ANOVA is still the most and venlafaxine ER. However, it is interesting to note commonly used statistical analysis tool for repeated some differences in tolerability and on efficacy measures measures in depression trials because of its simplicity observed in this analysis. For example, the proportion of and familiarity. However, mixed-effects models, the use completers (from the safety population) in the desvenla- of which has substantially increased over the last 10 years, faxine group (72%) was somewhat lower than in either may have important advantages over traditional methods the venlafaxine ER group (75–150 mg/day: 85%; 150– and may yield unbiased and more valid estimates 225 mg/day: 77%) or the placebo group (86%). This larger number of dropouts in the desvenlafaxine group mayreflect differences in tolerability; in particular, the rate ofnausea was higher in the desvenlafaxine group (38%) than The high placebo response observed in the two studies in either venlafaxine ER group (21 and 29% for the discussed herein reflects the increasing placebo response 75–150 mg/day and 150–225 mg/day dose ranges, respec- observed in many studies over recent years. It has been tively). This suggests that the higher dose range of des- estimated that the proportion of patients in studies, who venlafaxine used in these studies (i.e. 200–400 mg/ respond to placebo, has risen by approximately 7% per day), although safe and effective, may not be tolerated decade (Walsh et al., 2002). A high placebo response in a as well as lower doses. Likewise, the lower dose range of study makes it difficult, because of ceiling effects, to test venlafaxine ER had a lower discontinuation rate and may for efficacy, as a larger number of patients would be have been better tolerated than the higher dose range.
required for a valid comparison. Unless the power Differences between desvenlafaxine and venlafaxine ER calculations for the size of studies were constantly in the overall number and pattern of statistically revised upward to account for this difficulty, it would significant findings on secondary efficacy measures were be likely that the studies would be underpowered.
also observed in this pooled analysis. For example,although desvenlafaxine and both doses of venlafaxine Placebo response is a major issue in clinical trials for ER were statistically significant in the analysis of HAM- psychiatric disorders. The causes of a high placebo D17 and CGI-I scores, such consistency was not observed response in modern studies are many and varied and are in the responder analysis; rates of HAM-D17 response the subject of controversy. It is often claimed that the were not statistically significant with desvenlafaxine, but increased number of assessment visits and increased were so with both doses of venlafaxine ER. It is difficult nonspecific contact and support, which are part of current to draw conclusions based on this data because the study high contact trial practice, automatically increase the was not designed to compare the two active compounds.
response rate to both placebo and active treatment and Nevertheless, underlying differences in the effects of the raise the placebo response rate to a level where it is often difficult to distinguish from active treatment. Similarly,the use of many assessment instruments and the careful The ANCOVA, using the LOCF method, which tra- collection of AEs have the consequence of increasing ditionally has been used as a primary analysis in registration contact time with the treatment team, and this may need studies in EU and the US, uses the last recorded data to be more carefully controlled. Evidence also exists that point to replace the missing points for a participant failing the placebo response is higher in patients with mild- to complete the trial. This approach is believed to be the to-moderate depression and lower in those with more most conservative because it can reduce the apparent severe depression; therefore, the inclusion of patients efficacy by assigning high scores for medications that are with more severe depression has been recommended as a not well tolerated; however, recent comparisons of means of controlling the placebo response (Fava et al., different methods have demonstrated that this approach Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
International Clinical Psychopharmacology The treatment effect – the magnitude of change on the Lohr, Dieter Lohse Lutz, Ljiljana Moro, Peter Londborg, pivotal scale of an antidepressant compared with placebo Robert Bruce Lydiard, Eric Perreard, Liaudginas Erdvilas – is also likely to be affected by the raised placebo Radavicius, Paresh Ramjee, Paulis Revelis, Christine Reynaert, Robert Riesenberg, Norman Rosenthal, Eckart observed using an MMRM analysis is 2.3 points on the Ruther, Alexander Schulze, Bruno Scottez, Jeffrey Simon, HAM-D17, which compares well with the treatment Martelle Slabber, Jaroslaw Strzelec, Emilis Subata, Ivana effect of between 2 and 3 points that has been reported Timotijevic, Gerrit Christiaan Verster, Ryszard Wardenski, in positive placebo-controlled studies of the most recently licensed antidepressant, the SNRI duloxetine(Mallinckrodt et al., 2004). The treatment effect of Funding was provided by Wyeth Research Clinical Trials desvenlafaxine in this pooled analysis is also in line with Registry: NCT00087737 (US), NCT00090649 (EU).
that observed with the comparator venlafaxine ER,75–150 mg/day (2.4) and 150–225 mg/day (2.7), in the D. Lieberman has received grant/research support from individual studies. The effect size using a Cohen’s d AstraZeneca, Bristol Myers Squibb, Comentis, The Dalio measure was 0.41 for desvenlafaxine, 0.41 for venlafaxine Family Foundation, Eli Lilly, Epix, GlaxoSmithKline, ER 75–150 mg/day, and 0.46 for venlafaxine ER 150– McNeil, Ono, Predix, The Richard Lounsbery Founda- 225 mg/day. The clinical relevance of the significant tion, Sanofi Aventis, and Wyeth, and is on the speaker’s difference of the treatment effect observed with bureau of GlaxoSmithKline. S. Montgomery is a con- desvenlafaxine and venlafaxine ER is shown by the sultant to AstraZeneca, Bristol-Myers Squibb, Eli Lilly, significant advantage compared with placebo measured GlaxoSmithKline, Johnson & Johnson, Lundbeck, Merck, on the CGI-S and CGI-I scores, which represent the view Neurim, Pfizer, Pierre Fabre, Roche, Sanofi, Sepracor, of the independent clinician who is making a clinical Servier, Shire, and Wyeth. K. Tourian, K. Padmanabhan, judgment of the individual patient under double-blind and G. Rosas are employees of Wyeth Research, College- ville, Pennsylvania, USA. C. Brisard, J-M Germain, and B.
Pitrosky are employees of Wyeth Research, Paris, France.
ConclusionDesvenlafaxine was generally safe and well tolerated inthis population. Pooled analyses of the data with MMRM models demonstrate that desvenlafaxine was efficacious Alfinito PD, Huselton C, Chen X, Deecher DC (2006). Pharmacokinetic and in the treatment of MDD, despite the fact that the pharmacodynamic profiles of the novel serotonin and norepinephrine underpowered individual studies failed to reach statistical reuptake inhibitor desvenlafaxine succinate in ovariectomized Sprague-Dawley rats. Brain Res 1098:71–78.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (DSM-IV). 4th edition. Washington, DC: AmericanPsychiatric Association.
Bech P, Gram LF, Dein E, Jacobsen O, Vitger J, Bolwig TG (1975). Quantitative The clinical trials and analyses included in this report rating of depressive states. Acta Psychiatr Scand 51:161–170.
Clement EM, Odontiadis J, Franklin M (1998). Simultaneous measurement of were sponsored by Wyeth Research, Collegeville, Penn- venlafaxine and its major metabolite, oxydesmethylvenlafaxine, in human sylvania, USA. The authors thank Dorothy L. Tengler, plasma by high-performance liquid chromatography with coulometric detec- Publications and External Communications at Wyeth, for tion and utilisation of solid-phase extraction. J Chromatogr B Biomed SciAppl 705:303–308.
her professional writing and editorial assistance.
Deecher DC, Beyer CE, Johnston G, Bray J, Shah S, Abou-Gharbia M, et al.
(2006). Desvenlafaxine succinate: a new serotonin and norepinephrinereuptake inhibitor. J Pharmacol Exp Ther 318:657–665.
The authors also thank the following coinvestigators in DeLoach LJ, Higgins MS, Caplan AB, Stiff JL (1998). The visual analog scale in the desvenlafaxine 309 and 317 studies for their valuable the immediate postoperative period: intrasubject variability and correlation involvement: Raisa Andresina, Andris Arajs, Dolores with a numeric scale. Anesth Analg 86:102–106.
DeMartinis NA, Yeung PP, Entsuah R, Manley AL (2007). A double-blind, placebo- Backhaus, Bertrand Baranovsky, Bettina Bergtholdt, controlled study of the efficacy and safety of desvenlafaxine succinate Leszek Bidzan, Ralf Bodenschatz, Richard Brown, Wlod- in the treatment of major depressive disorder. J Clin Psychiatry 68: zimierz Chrzanowski, Johannes Coppens, Andrew Cutler, Faravelli C, Cosci F, Ciampelli M, Scarpato MA, Spiti R, Ricca V (2003). A self- Marek Cwiakala, Maciej Czerwinski, Didier Deroche, controlled, naturalistic study of selective serotonin reuptake inhibitors versus Vladimir Diligenski, Bernadette D’Souza, Christophe tricyclic antidepressants. Psychother Psychosom 72:95–101.
Dufour, David Louis Dunner, Louis Fabre, Michel Faure, Fava M, Uebelacker LA, Alpert JE, Nierenberg AA, Pava JA, Rosenbaum JF (1997). Major depressive subtypes and treatment response. Biol Psychiatry Pavo Filakovic, Michel Floris, Antoni Florkowski, Vera Folnegovic-Smalc, Peter Franz, Carllo Gagiano, Francis Gueorguieva R, Krystal JH (2004). Move over ANOVA: progress in analyzing repeated-measures data and its reflection in papers published in the Archives Gheysen, Janusz Janczewski, Mieczyslaw Janiszewski, of General Psychiatry. Arch Gen Psychiatry 61:310–317.
Riina Jents, Jacek Kacalak, Arifulla Khan, Louis Kirby, Guy W (1976). Clinical global impressions. Publication ADM 76–338. In: Guy W, editor. ECDEU assessment manual for psychopharmacology. Rockville, MD:US Department of Health, Education, and Welfare. pp. 217–222.
Louis Langeard, Philippe Leclercq, Pierre LeGoubey, Hamilton M (1960). A rating scale for depression. J Neurol Neurosurg Psychiatry Andres Lehmets, Joseph Lejeune, Vanda Liesiene, James Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Desvenlafaxine in the treatment of major depressive disorder Lieberman et al.
Liebowitz M, Yeung PP, Entsuah R (2007). A randomized, double-blind, placebo- Raskin A, Schulterbrandt J, Reatig N, McKeon JJ (1969). Replication of factors controlled trial of desvenlafaxine succinate in adult outpatients with major of psychopathology in interview, ward behavior and self-report ratings of depressive disorder. J Clin Psychiatry 68:1663–1672.
hospitalized depressives. J Nerv Ment Dis 148:87–98.
Lipman RS (1982). Differentiating anxiety and depression in anxiety disorders: Segal ZV, Pearson JL, Thase ME (2003). Challenges in preventing relapse in use of rating scales. Psychopharmacol Bull 18:69–77.
major depression. Report of a National Institute of Mental Health Workshop Mallinckrodt CH, Kaiser CJ, Watkin JG, Molenberghs G, Carroll RJ (2004). The on state of the science of relapse prevention in major depression. J Affect effect of correlation structure on treatment contrasts estimated from incom- plete clinical trial data with likelihood-based repeated measures compared Septien-Velez L, Pitrosky B, Padmanabhan SK, Germain J-M, Tourian KA (2007).
with last observation carried forward ANOVA. Clin Trials 1:477–489.
A randomized, double-blind, placebo-controlled trial of desvenlafaxine Molenberghs G, Thijs H, Jansen I, Beunckens C, Kenward MG, Mallinckrodt C, succinate in the treatment of major depressive disorder. Int Clin Psycho- et al. (2004). Analyzing incomplete longitudinal clinical trial data. Biostatistics Steffens DC, Krishnan KR, Helms MJ (1997). Are SSRIs better than TCAs? Montgomery SA, Asberg M (1979). A new depression scale designed to be Comparison of SSRIs and TCAs: a meta-analysis. Depress Anxiety 6: sensitive to change. Br J Psychiatry 134:382–389.
Muth EA, Moyer JA, Haskins JT, Andree TH, Husbands GEM (1991).
Thase ME, Entsuah AR, Rudolph RL (2001). Remission rates during treatment Biochemical, neurophysiological, and behavioral effects of Wy-45,233 and with venlafaxine or selective serotonin reuptake inhibitors. Br J Psychiatry other identified metabolites of the antidepressant venlafaxine. Drug Dev Res Walsh BT, Seidman SN, Sysko R, Gould M (2002). Placebo response in studies Nelson JC, Mazure CM, Jatlow PI, Bowers MB Jr, Price LH (2004). Combining of major depression: variable, substantial, and growing. J Am Med Assoc norepinephrine and serotonin reuptake inhibition mechanisms for treatment of depression: a double-blind, randomized study. Biol Psychiatry 55: World Health Organization (2001). The World Health Report: 2001–Mental Health: New Understanding, New Hope. Geneva: World Health Organization.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Source: http://www.dltmedwriting.com/samples/Desvenlafaxine_Tengler_Lieberman.pdf