Abstract
Objective:
To investigate if treatment outcome for severely depressed patients depends on their baseline level of anxiety.
Research design and methods:
Patients with a primary diagnosis of severe major depressive disorder (n = 459) were randomised to 24 weeks of double-blind treatment with escitalopram (20 mg) or paroxetine (40 mg). Post hoc analyses of efficacy in patients with a baseline HAM-A total score ≤20 (n = 171) or >20 (n = 280) were based on analysis of covariance (ANCOVA) (ITT, LOCF).
Results:
At week 24, the mean change from baseline in MADRS total scores was −24.2 for escitalopram-treated patients (n = 141) and −21.5 for paroxetine-treated patients (n = 139) (p < 0.05) in high baseline anxiety patients and the mean change from baseline in HAM-A total score was −17.4 (escitalopram) and −15.1 (paroxetine) (p < 0.05). When examining the proportion of complete remitters (CGI-S = 1) after 24 weeks of treatment, there was an increasing treatment difference as a function of baseline HAM-A total score in favour of escitalopram (ITT, LOCF). There was no treatment difference in the low baseline anxiety group. Significantly more patients (p < 0.01) withdrew from the paroxetine group (31%) than from the escitalopram group (17%), partly as the result of significantly more withdrawals due to AEs (p < 0.05). Incidence of AEs and withdrawals were not related to baseline anxiety and there were no significant differences in the incidence of individual AEs with escitalopram compared to paroxetine.
Limitations:
The post hoc nature of these analyses, the absence of placebo control group, and the requirement that patients should be suffering from severe depression, limit the generalisability of the results.
Conclusion:
Patients with severe depression together with comorbid anxiety symptoms responded significantly better to treatment with escitalopram 20 mg compared with paroxetine 40 mg. Contrary to paroxetine, escitalopram maintained its efficacy with increasing baseline anxiety levels.
Transparency
Declaration of funding
H. Lundbeck A/S, Denmark, funded the original randomised clinical trial, provided the analyses as needed and reviewed the manuscript.
Declaration of financial/other relationships
J.P.B. has disclosed that he is a consultant for, has received honoraria from and has conducted clinical research supported by Lundbeck. Biostatistician A.K.T.H. and E.W. have disclosed that they are fulltime employees of Lundbeck. A.H. declares no conflicts of interest.
Some peer reviewers receive honoraria from CMRO for their review work. The peer reviewers of this paper have disclosed that they have no relevant financial relationships.
Acknowledgment
The authors thank D.J. Simpson, Lundbeck, for technical assistance in the preparation of the manuscript.