Dr Ronald Pies questions the statements in regards to antidepressant response time from an October 2005 article in Psychiatric Times.
July 2006, Vol. XXIII, Issue 8
Dr Michael Posternak (Psychiatric Times, October 2005, page 34), in his article, How Quickly Do Antidepressants Begin to Work? provides intriguing and important new information suggesting that antidepressants may work far earlier than commonly believed. His research finds that antidepressant effects begin to separate from those of placebo within the first week of treatment. Perhaps so. But Dr Posternak's assertion that . . . antidepressants can either work quickly or they cannot requires careful parsing. Much turns on precisely what the key words in that statement actually mean. And, whereas it may be logically true that antidepressants either can or cannot work quickly, it may be empirically true that these drugs work over several different time frames, depending on a multitude of variables.
Indeed, I would suggest that there are at least 7 sources of heterogeneity in the actual time course of antidepressant response: (1) variation in drug pharmacodynamics and mechanism of action; (2) clinical heterogeneity within the diagnosis of major depressive disorder; (3) variability in the measured dimensions of clinical improvement (eg, psychic versus somatic symptoms of depression); (4) variability in plasma levels of antidepressants; (5) biologic variability in neuronal sensitivity to antidepressants, perhaps at the level of the gene; (6) sensitivity to early but secondary neurochemical effects of antidepressants; and (7) differences between early antidepressant responders (eg, within 3 to 5 days of drug initiation) and early early responders (eg, those who feel much better after just 1 day of treatment).
Each of these sources of variability would require a separate discussion, but they can be summarized as follows.
Pharmacodynamically, there is no reason to lump all antidepressants together and to assume that the time course will be the same for all; for example, paroxetine (Paxil) and sertraline (Zoloft)-inappropriately termed selective serotonin reuptake inhibitors-have prominent noradrenergic and dopaminergic effects, respectively.1
The DSM-IV criteria for major depressive disorder are so broad and elastic that patients with widely varying clinical pictures may be drawn into the same diagnostic net-for example, those with prominent guilt or suicidal ideation versus those with prominent weight loss and sleep disturbance.
Somatic signs of depression (eg, early morning awakening) may be more likely to respond quickly than intrapsychic dimensions, such as feelings of guilt or self-deprecation; thus, depending on how improvement is measured (eg, Hamilton Rating Scale versus Beck Depression Inventory), we may observe differing time courses of response.
It is by no means clear that all of the patients examined in studies of antidepressant response achieved adequate blood levels of antidepressants; hence, meta-analyses may be lumping together studies that achieved widely varying degrees of drug bioavailability.
Recent evidence suggests that antidepressants may work by inducing genes to produce neuronal growth factors, such as brain-derived neurotrophic factor; there is some basis for hypothesizing that individuals may differ in how inducible their genes are in this respect.
Anticholinergic and antihistaminic effects of some antidepressants may affect mood, sleep, and energy before the drug's primary mechanism of action kicks in, potentially skewing data on the time course of drug efficacy.
And, finally, there may be physiologic and psychological differences between patients who have an early response to antidepressants-within 3 to 5 days-and those who have a nearly instantaneous response (say, overnight). The first group may be benefiting from some aspect of the antidepressant's mechanism of action, whereas the nearinstantaneous responders may be experiencing a genuine placebo effect.
In summary, it seems likely that variation in the time course of antidepressant efficacy is a result of great heterogeneity among depressed patients and marked variability in our methods of assessing them. We should not be surprised by early, intermediate, or late responders to antidepressants.
Ronald Pies, MD
Dr Pies is clinical professor of psychiatry at Tufts University School of Medicine in Boston and Science Content Editor of Psychiatric Times.
1. Richelson E. Interactions of antidepressants with neurotransmitter transporters and receptors and their clinical relevance. J Clin Psychiatry. 2003;64(suppl 13):5-12.
I would like to thank Dr Pies for his thoughtful comments. In addition to the obvious heterogeneity of major depressive disorder itself, Dr Pies offers an erudite depiction of why there also may be great heterogeneity in the timing of response to therapy with antidepressant medications.
I fully agree with the conclusion that we should not be surprised by early, intermediate, or late responders to antidepressants. However, this has not been the standard teaching in the field. Psychiatric textbooks continue to teach that a true antidepressant response does not occur for 2 to 3 weeks after treatment is started. Dr Pies' letter, I would argue, only serves to further undermine the notion that all depressed patients can be lumped together and follow a uniform biochemical path to recovery.
The results of our meta-analysis clearly suggest that some patients do experience an almost immediate true drug response to antidepressant therapy. It would be of great interest to better understand which biochemical processes distinguish early from late responders. In our practice, we are continuing to accumulate data regarding the clinical profiles of early versus late responders, and we hope to publish this information within the next couple of years.
Michael A. Posternak, MD
Dr Posternak is assistant professor of psychiatry and human behavior at Brown University, in Providence, RI.