Effects of Initial Dosing Strategies of Duloxetine on Tolerability and Efficacy in Patients with Major Depressive Disorder

David Dunner, MD1 (sponsor); Susan Kornstein,MD2; Virgil Whitmyer, PhD3; Adam Meyers, MS3; Craig Mallinckrodt, PhD3; Madelaine Wohlreich, MD3; Millie Hollandbeck, BS3; John Greist, MD4

1) Center for Anxiety and Depression, Mercer Island, WA; 2) Virginia Commonwealth University, Richmond, VA; 3) Lilly Research Laboratories, Indianapolis, IN; 4) Healthcare Technology Systems, Inc., Madison, WI

ABSTRACTBackground: Patients often discontinue antidepressant treatment due to early side effects. Starting at a lower dose or taking the medication with food is often recommended to mitigate initial adverse events, but these strategies have not been well studied. Duloxetine is a serotonin-norepinephrine reuptake inhibitor with an efficacious (recommended) dose of 60mg once daily (QD). A previous open-label duloxetine study suggested that starting duloxetine at 30mg QD for one week followed by escalation to 60mg QD may lessen the risk of initial nausea with only a short-lived impact on efficacy. This study compared starting doses of duloxetine taken with or without food.

Methods: This double-blind, parallel design trial was conducted in adult outpatients with major depressive disorder (MDD). Patients were randomized in a 3 x 2 complete factorial arrangement to one of three starting dose groups: 30mg once daily in the morning (QAM; n=219), 30mg twice daily (BID; n=213), or 60mg QAM (n=215) and to one of two food groups: by instruction to take study drug with food or not within one hour of eating. After one week on the starting dose, all patients were dosed at 60mg QD for the remaining five weeks of treatment. The primary objective of the study was to compare the rate of treatment emergent nausea in the 30mg QAM group versus the 60mg QAM group based on item 112 (nausea) of the Association for Methodology and Documentation in Psychiatry adverse event scale (AMDP-5). A key secondary objective was to evaluate mean changes on AMDP-5 item 112 using an analysis that included dose group, food group, and their interaction. Other secondary objectives included mean changes on an a priori determined common adverse events list and discontinuations due to adverse events. Efficacy was primarily evaluated by the 17-item Hamilton Depression Rating Scale (HAMD) and the Inventory of Depressive Symptoms-Clinician Rated (IDS).

Results: No significant differences were found between starting doses of 30mg QAM and 60mg QAM on the primary analysis of rate of treatment-emergent nausea. However, on the secondary mean change analysis, both the main effect of food group and the starting dose group by food group interaction were significant at week 1. These results differed from the primary analysis because the mean change analysis assessed changes in both rate and severity of nausea. Further, there was a main effect of food - taking duloxetine with food reduced (improved) initial nausea—with the greatest benefit of food among those patients who started at 60 mg QAM. When taking duloxetine without food, patients starting at 30mg QAM had improved nausea compared with those at 60mg QD. In patients taking duloxetine without food discontinuation rates due to adverse events were 3.6%, 14.0%, 10.2% versus those taking it with food at 5.4%, 7.5%, and 7.4% for 30mg QAM, 30mg BID, and 60mg QAM respectively. All starting dose groups showed significant baseline to endpoint improvements, as measured by mean changes in the HAMD. However, patients initially dosed at 60mg QAM showed significantly greater improvement at weeks 1and 2 (IDS) and 2 (HAMD) than those initially dosed at 30mg QAM or 30mg BID. For the remaining 4 weeks of treatment, mean change did not differ among initial dose groups.

Conclusions: Taking duloxetine with food or starting at 30mg QAM appeared to improve initial tolerability. Starting 30mg BID did not improve tolerability compared to 60mg QAM. The starting dose of 30mg QAM in the first week produced a transient disadvantage in efficacy compared to a starting dose of 60mg QAM.

BACKGROUND• Major depressive disorder (MDD) has a lifetime prevalence ranging from 10% to 25% in females and 5% to 12% in males (APA 2000).

• Treatment-emergent nausea associated with initial antidepressant treatment is one reason for treatment discontinuation and is reported at a rate of 20% for selective serotonin reuptake inhibitors (SSRIs), with rates as high as 31% for venlafaxine, a dual-reuptake inhibitor of serotonin and norepinephrine (SNRI) (PDR 2003).

• Duloxetine is a potent dual reuptake inhibitor of serotonin and norepinephrine; exhibits a low affinity for most neurotransmitter receptors (Wong and Bymaster 2002); thus, suggesting a favorable side effect potential.

• In trials starting duloxetine at the effective treatment dose for MDD of 60 mg QD, the nausea rate was 38%, which appeared to be short-lived and dose related.

• A recent open-label study further suggested that taking duloxetine 30 mg QD for one week followed by escalation to doses of 60 mg or higher may lessen the risk of adverse events with only a short-lived impact on efficacy compared with starting at 60 mg QD.

• Starting duloxetine at lower doses or taking duloxetine with food is often recommended to mitigate initial adverse events but has not been well-studied. This study aimed to examine these dosing strategies.

OBJECTIVESPrimary objective

• Compared the incidence of treatment-emergent nausea for patients initially dosed at duloxetine 30 mg QAM, versus duloxetine 60 mg QAM.

Secondary objectives

• Examined effects of dose and food on tolerability and efficacy in a 3 x 2 factorial design. Starting Doses: 30 mg QAM; 30 mg BID; and 60 mg QAM compared with taking starting dose with or without food.

• Compared tolerability and efficacy after 1 week at the starting dose and at the end of the treatment (5 weeks).

METHODS• Double-blind parallel design trial conducted in male and female adult outpatients (18 years or older) with MDD as defined by DSM-IV-TR.

• Patients were randomized in a 3 x 2 complete factorial arrangement to:

Starting dose groups: 30mg QAM (n=219); 30mg BID (n=213); 60mg QAM (n=215)

Food groups: by instruction to take study drug with food or not within one hour of eating (without food).

• After 1 week on the starting dose, all patients were dosed at 60mg once daily (QD) for the remaining 5 weeks of treatment.

• Assessment of nausea was done by the AMPD-5 and included:

Mean change on AMDP-5 item 112 (nausea)

• The “common adverse events score” is the mean of items from the AMDP-5 (defined a priori) that are commonly associated with duloxetine. These items include: nausea, vomiting, dry mouth, constipation, mean of insomnia items, drowsiness, increased perspiration, and decreased appetite.

Funding provided by Eli Lilly and Company

CONCLUSIONS• Overall, duloxetine was efficacious and well-tolerated regardless of starting dose.

• In this study, instructing patients to take duloxetine with food improved initial tolerability, particularly in patients started at 60mg-QAM.

• Starting patients at 30mg-QAM improved tolerability compared to starting doses of 30mg-BID and 60mg-QAM.

• Starting patients at 30mg-QAM resulted in a transitory delay in efficacy which was not significant after Week 2.

• Remission rates at the end of the treatment period were not different among the 3 starting doses and were 39.6% (30mg QAM), 35.9% (30mg BID), and 42.1% (60mg QAM).

REFERENCES[PDR] Physicians’ Desk Reference. 2003. Montvale (NJ): Medical Economics Data Production Co.

Wong DT, Bymaster FP. 2002. Dual serotonin and noradrenaline uptake inhibitor class of antidepressants potential for greater efficacy or just hype? Prog Drug Res 58:169-222.

APA. 2000. Diagnostic and Statistical Manual of Mental Disorders Fourth Edition - Text Revision. Washington DC.

HMDR Study Design – Acute Phase

Demographics

Incidence of Treatment-Emergent Nausea at Week 1 and Acute Phase by Dose

METHODS continued• Efficacy was primarily evaluated by the 17-item HAMD and the 30-item Inventory of Depressive Symptoms Clinician Rated (IDS-30) total score.

Statistical Methods

• All patients with a baseline and at least one post-baseline measure were included in the analysis.

• Fisher’s exact test was used to assess the equality of dose groups in the incidence of treatment-emergent nausea based on changes in AMDP-5 item 112 at Week 1. The primary analysis was the contrast between the 30 mg QAM and 60 mg QAM dose groups.

• HAMD and IDS-30 mean changes from baseline to all post baseline visits were analyzed using a restricted maximum likelihood (REML)-based repeated measures approach (MMRM). Analyses included the fixed, categorical effects of dose group, investigator, visit, and dose group-by-visit interaction, as well as the continuous, fixed covariate of baseline score. AMDP-5 outcomes were analyzed as described above with the addition of food group, food group-by-visit interaction, and food group-by-dose group-by-visit interaction as categorical effects to the model.

DLX 30 mg QAM (n=111)

DLX 60 mg QAM (n= 108

DLX 30 mg BID (n= 107

1 week

DLX 30 mg QAM (n=106

DLX 60 mg QAM (n=108

DLX 30 mg BID (n=107

“Take with food

“Do not take within an hour of eating

8 week extension

Week 0 Week 1 Week 6

After 1 week all patients: 60mg QD dose Duloxetine

DLX 30 mg QAM (n=111)

DLX 60 mg QAM (n= 108)

DLX 30 mg BID (n= 107)

1 week

Acute phase = 6 weeks

DLX 30 mg QAM (n=106)

DLX 60 mg QAM (n=108)

DLX 30 mg BID (n=107)

“Take with food ”

“Do not take within an hour of eating”

8 week extension

Week 0 Week 1 Week 6

After 1 week all patients: 60mg QD dose Duloxetine

CharacteristicDLX 30mg QAM

(N=219)

DLX 30mg BID

(N=213)

DLX 60mg QAM

(N=215)

Gender, N (%) Female 136 (62.1) 145 (68.1) 134 (62.3)

Age, y, mean 42.2 42.8 43.9

Age, y, range 18.9 - 80.1 18.7 - 82.7 18.6 – 77.5

Ethnicity, n (%)

Caucasian 169 (77.2) 162 (76.1) 174 (80.9)

African descent 28 (12.8) 26 (12.2) 15 (7.0)

Hispanic 15 (6.8) 20 (9.4) 19 (8.8)

Others 7 (3.2) 5 (2.3) 7 (3.3)

HAMD-17 total, mean 21.6 21.7 21.2

IDS-30 total, mean 35.5 35.9 35.1

CGI-S, mean 4.4 4.3 4.3

Mean Nausea Score Over Acute Phase by Dose

Nausea Score – Change Among Patients Reporting New Nausea at Week 1

2327 29

95 3

0

10

20

30

40

50

Per

cen

t

Pooled Across Food Groups

Week 1

Weeks 2-6

30 mg QAM 30 mg BID 60 mg QAM0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 1 2 3 4 5 6Weeks

Me

an

Sc

ore

30 mg QAMStarting Dose

30 mg BID60 mg QAMStarting Dose

Imp

rov

em

en

t

*p = .04 (60mg QAM starting dose vs30 mg QAM starting dose)

*

After Week 1 – all patients were dosed at 60 mg QD

Incidence of Treatment-Emergent Nausea at Week 1 by Food and Dose

26

30

33

19

24 24

0

10

20

30

40

50

Per

cen

t

Without Food With Food

30 mg QAM

30 mg QAM

30 mg BID

30 mg BID

60 mg QAM

60 mg QAM

There were no differences within or between dose and food groups in the incidence of treatment-emergent nausea

Incidence of Common Adverse Events (AMDP-5) at Week 1 by Dose

Mean Change in HAMD17 Total Score Over Acute Phase

17

22

14 12

1915

8

5

16

3

4

7

2

3

3

4

1 2

0

5

10

15

20

25

30

35

30 mg QAM

30 mg BID

60 mg QAM

30 mg QAM

30 mg BID

60 mg QAM

Without food

Pe

rce

nt

With Food1 unit of change in severity: 0 -1, 1-2, or 2-3

2 units of change of severity: 0-2 or 1-3

3 units of change of severity: 0-3

Nausea Severity: 0 = no nausea; 1 = mild; 2 = moderate; 3 = severe

Mean Change of Nausea Score at Week 1 by Food and Dose

0.26 0.28

0.50

0.20 0.15 0.18

0

1

Me

an

Ch

an

ge

fro

m B

as

eli

ne

Without Food With Food

* p=.006 (60 QAM w/out food vs

60 QAM w/food)

p = .01 for main effect of food

30 mg QAM

30 mg QAM

30 mg BID

30 mg BID

60 mg QAM

60 mg QAM

p = .03 (30 QAM vs60 QAM)

p =.06 (30 BID vs60 QAM)

*

Mean Change of Common Adverse Event Score at Week 1 By Food and Dose

0

0.870.82

0.50

0.37

0.24

0

1

Me

an

Ch

an

ge

fro

m B

as

eli

ne

Without Food With Food

30 mg QAM

30 mg QAM

30 mg BID

30 mg BID

60 mg QAM

60 mg QAM

p = .35 for main effect of foodp=.02

p=.01

Mean Common Adverse Event Score Over Acute Phase by Food and Dose

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

0 1 2 3 4 5 6Weeks

Me

an

Sc

ore

30 mg QAM with Food 30 mg QAM w/o Food 30 mg BID with Food

30 mg BID w/o Food 60 mg QAM w/ food 60 mg QAM w/o Food

Imp

rov

em

en

t

1-week at starting doses

All groups at 60 mg QD

End of Placebo lead in

Baseline Scores:30mg QD w/food 4.6 30mg QD w/out food 4.7

30mg BID w/food 4.930mg BID w/out food 4.8

60mg QD w/food 4.5 60mg QD w/out food 4.5

Discontinuation Due to Adverse Event in Acute Phase by Dose and Food

0

5

10

15

20

25

30

35

An

yIn

som

nia

Gas

tric

Eve

nts

Dro

wsi

nes

s

Dry

Mo

uth

Incr

ease

dP

ersp

irat

ion

Diz

zin

ess

Co

nst

ipat

ion

Pe

rce

nt

Inc

ide

nc

e

Duloxetine30 mg QAM

Duloxetine30 mg BID

Duloxetine60 mg QAM

*p =.04 (30mg QD vs 30mg BID)Any Insomnia - Mean of difficulty falling asleep, interrupted sleep, shortened sleep, and early waking; Gastric Events - Mean of nausea and vomiting

1.8

7.0

5.1

2.7

3.83.7

0

5

10

Pe

rce

nt

of

Pa

tie

nts

Without Food With Food

30 mg QAM

30 mg QAM

30 mg

BID

30 mg BID

60 mg QAM

60 mg QAM

p = .07 (30 QAM vs 60 QAM w/o food)

p = .01 (30 QAM vs 30 BID w/o food)

n=15 n=8n=6n=11n= 4

n=8

At Week 1, there were no statistically significant differences between starting doses within each food group in discontinuation rates due to adverse events

Mean Change of IDS-30 Total Score Over Acute Phase

-10

-9

-8

-7

-6

-5

-4

-3

-2

-1

00 1 2 3 4 5 6

Weeks

Me

an

Ch

an

ge

fro

m B

as

eli

ne

Duloxetine30 mg QAM

Duloxetine30 mg BID

Duloxetine60 mg QAM

Imp

rov

em

en

t

**

*

*p =.01 (60 QAM vs 30 QAM)

**p< .001 (60 QAM vs 30 BID)

Baseline Mean Score = 17.55 (after placebo lead in)Mean change in HAMD17 Total Score at endpoint was statistically significant for each starting dose group

-15

-13

-11

-9

-7

-5

-3

-1

0 1 2 3 4 5 6

Weeks

Me

an

Ch

an

ge

fro

m B

as

eli

ne

Duloxetine30 mg QAM

Duloxetine30 mg BID

Duloxetine60 mg QAM

Imp

rov

em

en

t

**

*

**

*

Week 1: * p=.05 (60 QAM vs 30 QAM)**p < .005 (60 QAM vs 30 BID)

Week 2: * p=.03 (60 QAM vs 30 QAM)**p.=.001 (60 QAM vs 30 BID)Baseline Mean Score = 29.20 (after placebo lead in)

Mean change in IDS-30 Total Score at endpoint was statistically significant for each starting dose group.

*