LOS ANGELES – Antidepressant medications have a robust effect in older patients with a long duration of major depressive disorder that is at least moderately severe, according to findings from a meta-analysis of data from seven trials.
Since patients with a long duration of disease are also at increased risk for recurrence, and since the efficacy of antidepressant medications for relapse prevention is well established, antidepressant treatment should be considered in these patients, Dr. J. Craig Nelson said at the annual meeting of the American Association for Geriatric Psychiatry.
©Dundanim/shutterstock.com
Antidepressants can be very effective at treating severe depression and anxiety in seniors, said Dr. J. Craig Nelson.
Several prior studies have demonstrated that all classes of antidepressants generally are better than placebo for the treatment of depression in older adults, and that they have similar effectiveness, but the effects are modest, said Dr. Nelson, a professor of psychiatry at the University of California, San Francisco.
For example, in a 2008 meta-analysis of data from 10 studies that included adults aged 60 years and older with major depressive disorder, the response rates in the antidepressant and placebo groups were 44.4% vs. 34.7%, respectively (odds ratio, 1.40), and the remission rates were 34.6% vs. 26.5%, respectively (OR, 1.27), he said (Am. J. Geriatr. Psychiatry 2008;16:558-67).
The findings raised the question of whether moderators associated with a more robust response could be identified, he said, noting that several open studies have suggested that those with high levels of anxiety have lower responses to drug treatment, and several other moderators, such as disease severity and recurrent depression, appear to predict worse outcomes.
To look more closely at such moderators, Dr. Nelson and his colleagues revisited the 10 studies included in their 2008 meta-analysis, and obtained trial-level data from 8 of the studies for which complete information was available. They found no difference in outcomes between anxious and nonanxious depressed patients treated with antidepressants (Int. J. Geriatr. Psychiatry 2009;24:539-44).
The investigators decided to take an even closer look at possible moderators of response by performing a meta-analysis using patient-level data from seven of those trials for which complete information was available. These trials included 2,283 patients (mean age, 71.4 years), a mean duration of major depressive disorder of 11.8 years, and a mean baseline Hamilton Depression Rating Scale (HDRS) score of 21.5. About two-thirds (64.6%) were women, and nearly three-fourths (73.9%) had recurrent depression.
The findings, which were pending publication in the American Journal of Psychiatry at the time of Dr. Nelson’s presentation, showed significant linearity for the relationship between duration of illness and response in the placebo group, but not in the drug group (z scores, –3.81 and –0.92, respectively), and baseline depression severity was significantly associated with response in the drug group but not in the placebo group (z scores, 3.40 and –0.40, respectively).
On multivariate analysis, the strongest independent predictor of response was duration of illness, followed by severity of depression. An interaction between severity and duration also was noted.
"It turned out that among patients with longer duration of depression, severity did have a more significant relationship with drug-placebo difference. Among patients with a long duration of depression, patients with severity of at least moderate intensity – a Hamilton score of 21 or greater – showed the greatest drug effects," Dr. Nelson said.
Among patients with depression duration of less than 10 years and an HDRS score of less than 21, the response rates were 46.3% and 41.5% in the drug treatment vs. placebo groups, respectively, with a number needed to treat of 21; among patients with depression duration greater than 10 years and an HDRS score of 21 or greater, the response rates were 58% and 31.4% in the groups, respectively, with a number needed to treat of 4.
Of note, while disease duration and severity appear to identify patients who are more drug responsive, they also identify patients who have a pretty good response to placebo, he added.
That is, patients with a shorter duration of depression and less severe disease tend to have a good placebo response. Such a response is seen in 40%-50% of such patients.
"But of course, this is not just placebo. These are patients who are being seen, usually, on a weekly basis for the first month and then maybe every other week for the remainder of the trial," he explained, noting that the clinical management remains an important aspect of care.
Given that many trials include extensive clinical management along with drug treatment, it cannot be concluded that drug treatment without such management would lead to improvement, he said.