Duloxetine in the Treatment of Major Depressive Disorder: RESULTS

Patient Characteristics

Baseline patient demographics are summarized in Table 2. African-American patients were significantly younger (p=0.001), and had significantly higher mean body weight (p=0.049), when compared with Caucasian patients. There were no significant differences in psychiatric profile at baseline (Table 2). With regard to psychiatric history, African-American patients had a significantly higher mean age at onset of depression compared with Caucasians (29.7 years vs. 26.5 years; p=0.016). There were no significant between-group differences in any other aspect of psychiatric history (including duration of current episode, number of previous episodes, or proportion of patients with atypical or melancholic features).

Table 2. Patient baseline demographics and psychiatric history (all studies)

	African American (N=128)	Caucasian (N=1342)	P Value
Gender, n (%) Female	90 (70.3)	863 (64.3)	0.208
Age (Years) Mean (SD)	38.6 (12.0)	42.3 (13.1)	0.001
Age Range Min-Max	18-71	18-82
Weight (kg) Mean (SD)	87.1 (24.3)	83.4 (21.2)	0.049
HAMD i7 Total Score Mean (SD) 21.7 (4.0)		21.1 (4.1)	0.529
CGI-Sevehty Mean (SD)	4.18 (0.46)	4.30 (0.55)	0.130
VAS Overall Pain Mean (SD)	31.4 (23.0)	31.2 (25.1)	0.766

Efficacy

All studies. Analyses of efficacy data from all seven studies are presented in Table 3. Treatment-by-ethnicity interactions were not statistically significant, indicating that the magnitude of duloxetine’s treatment effects did not differ significantly between African-American and Caucasian patients. Advantages for duloxetine over placebo in HAMDn, CGI-S and PGI-I measures were highly significant (p<0.001) among Caucasian patients. In the substantially smaller group of African-American patients, duloxetine-placebo differences did not achieve statistical significance. The effect size for mean change in HAMD17 total score in African-American patients was 0.17, compared with 0.24 in Caucasians. Effect sizes for mean change in CGI-S score were 0.16 (African Americans) versus 0.22 (Caucasians).

Table 3. Summary of efficacy measures

Mean Change (SD)
HAMDi7 Total Score	Placebo	Duloxetine	P Value0	P Value6
All Studies Caucasians (n= 1,300) African Americans (n=123)	-5.99 (7.44) -6.36 (6.56)	-7.72 (7.07) -7.66 (8.81)	O.001 0.741	0.328
Positive Studies Caucasians (n=737) African Americans (n=59)	-5.58 (7.19) -7.19 (6.20)	-8.04 (7.04) -10.29 (8.02)	O.001 0.092	0.985
CGI-Severity All Studies Caucasians (n=l,301) African Americans (n=123)	-1.03 (1.25) -1.04 (1.08)	-1.31 (1.24) -1.24 (1.36)	O.001 0.671	0.298
Positive Studies Caucasians (n=738) African Americans (n=59)	-0.96 (1.23) -1.10 (1.01)	-1.31 (1.23) -1.75 (1.21)	O.001 0.050	0.629
PGI-lmprovement All Studies Caucasians (n=l,301) African Americans (n=123)	3.15 (1.29) 2.77 (1.08)	2.77 (1.30) 2.75 (1.22)	O.001 0.786	0.130
Positive Studies Caucasians (n=738) 3.29(1.33) 2.79(1.32) O.001 0.314 African Americans (n=59) 2.68 (1.01) 2.54 (1.26) 0.631 CGI: Clinical Global Impression; HAMD: Hamilton Rating Scale for Depression; PGI: Patient Global Impression; a: p value for duloxetine medication versus placebo; b: p value for treatment-by-ethnicity interaction

Positive studies. Analyses of pooled efficacy data from the four positive studies yielded results similar to those from the pooling of all studies (Table 3). Treatment-by-ethnicity interactions for HAMD17, CGI-S and PGI-I outcomes were not statistically significant, indicating that the magnitude of duloxetine’s treatment effects were similar in African-American and Caucasian patients. In Caucasian patients, duloxetine’s advantage over placebo was highly significant (p<0.001) for all assessed outcomes, while in African-American patients the improvement in mean CGI-S score was significantly greater for duloxetine-treated patients compared with placebo (p=0.050). The effect size for mean change in HAMD17 total score in African-American patients was 0.44, compared with 0.35 in Caucasians. Effect sizes for mean change in CGI-S score were 0.59 (African Americans) versus 0.28 (Caucasians).

Table 4. Summary of efficacy measures (African-American patients, focus studies)

Mean Change (SE)
	Duloxetine 30 mg QD (n=16)	Placebo (n=16)	P Value
H AMD и Total Score	-12.90 (1.93)	-9.23 (2.05)	0.192
HAMD i7 Subscales Core Maier Anxiety Retardation Sleep	-6.54 (0.93) -7.53 (1.09) -3.43 (0.66) -4.91 (0.79) -2.00 (0.49)	-3.37 (1.01) -4.41 (1.17) -2.43 (0.71) -3.04 (0.85) -1.96 (0.53)	0.021 0.049 0.302 0.105 0.964
HAMD и Items Item 1 (depressed mood) Item 7 (work and activities)	-1.92 (0.31) -1.98 (0.31)	-1.25 (0.34) -1.09 (0.34)	0.145 0.052
CGI-Severity -2.25(0.33) -1.41(0.35) 0.083 PGI-lmprovementa 2.10(0.34) 2.61(0.36) 0.300 QLDSb -14.64(10.00) -4.50(6.88) 0.231 CGI: Clinical Global Impression; HAMD: Hamilton Rating Scale for Depression; PGI: Patient Global Impression; QLDS: Quality of Life in Depression Scale; a: mean score; b: mean change (SD)

Focus studies. In the two focus studies (1 and 2), the magnitude of duloxetine’s treatment advantage over placebo in African-American patients was similar to that observed in Caucasian patients for HAMD17, CGI-S and PGI-I assessment measures. However, the small sample size of the African-American group (n=16 for generic cymbalta; n=16 for placebo) resulted in few outcomes achieving statistical significance. Only the two HAMD17 subscales assessing core emotional symptoms of depression (core and Maier subscales) demonstrated significant advantages for duloxetine-treated patients when compared with placebo (Table 4).

The response rate among duloxetine-treated African-American patients in the two focus studies was 50%, compared with 38% for patients receiving placebo (p=0.722). Remission rates were 44% vs. 25% for African-American patients receiving generic duloxetine and placebo, respectively (p=0.458). In comparison, response rates among Caucasian patients in the two focus studies were 47% vs. 29% for duloxetine and placebo, respectively (pO.OOl), while remission rates were 30% vs. 20% for Caucasian patients receiving duloxetine and placebo, respectively (p=0.039).

Table 5. Summary of VAS pain measures0 (African-American patients, focus studies)

Mean Change (Percentage Change)
Duloxetine 30 mg		QD (n=	=16)	Placebo (n=16)	P Value
Visual Analog Scale
Overall	-0.70	(-3.3)		6.16 (29.2)	0.192
Headache	-7.66 (	-35.8)		-3.03 (-14.1)	0.310
Back pain	-4.02 \|	-17.3)		-0.45 (-1.9)	0.720
Shoulder pain	2.41 \|	18.2)		5.14 (38.8)	0.599
Interference with daily activities	-0.18	(-1.2)		6.02 (40.0)	0.367
Time in pain while awake	0.81	(3.5)		5.30 (22.8)	0.517
a: Main effect of treatment (pooling all visits)

In analyses focusing on the main effect of treatment for VAS pain measures, duloxetine-treated African-American patients demonstrated greater improvement compared with placebo on all six assessed outcomes (Table 5). However, these advantages did not reach statistical significance. Effect sizes for the six pain outcomes ranged from 0.14 (shoulder pain) to 0.28 (overall pain).