ANTIDEPRESSANTS, SUICIDALITY AND ACADEMIC INTEGRITY Earlier this year, the American Journal of Psychiatry ran an editorial, asking: "Is the FDA Warning About Antidepressants Wrong?" Since 2004, the Food and Drug Administration (FDA) has warned that, suicidal behaviour may emerge after antidepressant treatment is begun - but without specifically suggesting that antidepressant drugs were the cause.
The Am J Psych decided that: "reconsideration of the FDA warning seems warranted". This was on the basis of a major, NIH-funded, epidemiological study, published in the same issue that concluded, "Available data do not indicate a significant increase in risk of suicide or serious suicide attempt after starting treatment with newer antidepressant drugs". Much publicity followed: "Antidepressants Seem to Cut Suicide Risk in Teenagers and Adults, Study says", according to the New York Times.
After examining thousands of records in a Seattle-based, "consumer-governed", non-profit health plan, the researchers reported that "suicide attempts decreased by over 60% in the first month of treatment". As suicide attempts typically bring people into contact with doctors who then usually prescribe an antidepressant drug, this study just quantified the obvious: levels of suicidal behaviour tend to fall when people get help.
The study also looked at the risks of suicidal behaviour with patients on drugs. We now know that the risks for depressed patients treated with paroxetine might be of the order of 1:100, and six times higher than for patients who see doctors but get a placebo. But this study, by Dr. Gregory E. Simon and colleagues, had no placebo group, and they found that: "The risk of suicide during acute-phase antidepressant treatment is approximately one in 3,000 treatment episodes, and risk of serious suicide attempt is approximately one in 1,000. Available data do not indicate a significant increase in risk of suicide or serious suicide attempt after starting treatment with newer antidepressant drugs"(American Journal of Psychiatry 163:34A, January 2006, doi: 10.1176/appi.ajp.163.1.A34).
And what of the differences in suicidal behaviour at the start of treatment and beyond? Here the evidence was much less certain, in spite of the trumpeting that emerged through a combination of spin and sequential whispers - and so a correspondence ensued. Make up your own mind on the basis of your reading of the Simon et al study and the rather one-sided e-mail correspondence that follows. Our view would be that lack of accountability is a sign of weakness and damaging to patients interests and that, in the absence of explanation, a retraction would seem due. A note to this effect has now been sent to the editors of the Am J Psych; you can leave feedback for them here.10 January 2006
I'm sorry to trouble you, but I wonder if you could help with this query arising from your recent paper (Am J Psychiatry 163:34A, January 2006) on Suicide Risk During Antidepressant Treatment.
In the text immediately preceding Figure 5, you and your colleagues noted that: "Suicide attempts during the first month of treatment were relatively evenly distributed throughout the month", (as Figure 5 shows). However, in paragraph two of the Discussion, you also noted that, "In agreement with Jick and colleagues (15), we found a significantly higher risk of suicide attempts in the first week of antidepressant treatment than in subsequent weeks".
Forgive if I've missed something, but I am unable to reconcile these two statements on the basis of the data and interpretations presented. I would be most grateful if you could clarify the position and put me right. Thank you for your attention and I look forward to hearing from you.
Social Audit Ltd
11 January 2006
Dear Dr Simon,
Thank you for your mail, but your response raised further questions in my mind. If, as you say, the risk of serious suicide attempts in week 1 was not higher than in weeks 2, 3, and 4, it would seem to follow that the risks in first four weeks of treatment were greater than in subsequent weeks (from week 5 onward). Does this honestly justify the headline of your press release, "Suicide risk does not increase when adults start using antidepressants"?
Your paper also states that, "Among those treated with newer antidepressants (Figure 6) risk in the first month of treatment was not significantly higher than in months 2 - 6". You attribute the increased risk in the first month of treatment to treatment with older antidepressant medications yet you also acknowledge that, "older drugs were more often prescribed at the beginning of the study period, (1992 to June 2003) when the rates of hospitalization were generally higher".
Your paper noted that when you correct for this anomaly, "the risk of suicide death and the risk of suicide attempt were not significantly lower in patients treated with new drugs". Why was this seemingly important qualification not mentioned in your press release? "The study also found that the risk of suicidal behavior after starting 10 newer antidepressant medications is less than the risk posed by older medications".
I appreciate and share your general concern that people should not be deterred from taking antidepressants because they over-estimate the risk of drug-induced suicidal behaviour but with these qualifications.
1. The headline in your press release, "Long-term, population-based study challenges FDA advisory" seemed unwarranted. It failed to acknowledge the critical lack of any placebo or control group in your study and, as you acknowledge in the small print of the paper, "our data certainly do not exclude the possibility that antidepressant drugs may precipitate increased suicidal ideation or suicide attempts in a sub-group of vulnerable individuals", and "an observational study such as this one can neither clearly establish nor clearly refute a causal relationship between antidepressant use and risk of suicide".
2. In a paper on "suicide risk during antidepressant treatment" (emphasis added), it did not seem appropriate to use as a prominent major benchmark the inevitably very high rate of suicidal behaviour that occurred in the month before drug treatment. That measure appears to be a cockshy: suicidality emerging before (and leading to) treatment clearly has nothing to do with the propensity of antidepressants to cause suicidal behaviour.
3. In the UK, and probably also in the US, about two-thirds of all prescriptions for SSRIs are for mild depression (only 3% are for severe depression), yet there is still no good evidence that SSRIs are more effective than placebo in treating cases of mild depression. When there can be no expectation of significant benefit, what justification could there be for prescribing these drugs for people with mild depression, exposing them to significant risks? Did your study explore this, and do the prescribing practices recommended in the Group Health Cooperative take account of severity of depression?
I apologise in advance if my understandings seem to you limited or incomplete but considering the available evidence, alongside the interpretations and emphases added, I feel far from persuaded that no evidence of risk = evidence of no risk. I would welcome your comments on these further points.
11 January 2006
Could you please identify more clearly the purpose of your questions. Is this merely for our own information and interest?
11 January 2006
Dear Dr Simon,
I hope that this deals satisfactorily with your query and look forward to hearing from you.Sincerely,
It's been some weeks since I wrote, I hope explaining the reasons for the questions I raised about your paper. I also sent you by airmail a copy of the book, Medicines out of Control? and I hope this arrived safely.
I am now writing to ask if and when I might expect a reply from you. I would appreciate your response on the points I raised.
19 May 2006
Ive been meaning to write to you for some time and the recent publicity about paroxetine provides a good opportunity to do so. This is to advise you that I shall shortly be posting a note on the Social Audit website about our correspondence earlier this year. You will see that the emphasis is on the need for accountability, rather than lack of courtesy. That seemed appropriate, given my views on the quality of your paper, and your conviction that there was never any need to defend it.
I receive approximately 300 email messages a day. I simply can't respond at length to them all. If you would choose to raise questions about our paper, the usual method in the scientific community would be to write a letter to the editor of the journal publishing the paper. Please feel free to do that. I have responded to all questions raised through public channels. I do not have the time to repond to all private correspondence. If you value accountability, then you would certainly want all correspondence (including this message) to be public.
Greg Simon21 May 2006
Thank you for your mail. I appreciate that you would not be able to respond at length to 300 emails/day, but suspect you need a good spam filter. Thank you for telling me about protocol in the scientific community: I meant no offence when writing to first check with you about some of the queries I had about your paper, before writing to the journal concerned. However, I have now done so: see attached. I shall of course include our correspondence when I post to our website, and if you would like to comment further, of course you have the opportunity to do so.