Back to 3.7

3.8 Risks of long-term use

Evidence relating to possible adverse effects with long-term use is sparse but there is little evidence of concern, probably because the risks of non-intervention are considered far greater. In addition, patients who experience the worst unwanted effects tend either to quit early on, or develop tolerance to them if they persist.

The relatively few studies of long-term use mainly focus on efficacy and most last for one year. Research into long-term adverse effects would be complicated, expensive and hard to fund. Also bearing in mind that SSRIs have not yet been used long for enough to be sure of their effects, the risks associated with indefinite use can only be guessed at. Unexpected problems and the risk of insidious harm cannot be ruled out. They could become apparent only well into the future, as they have with the tricyclics and many other drugs:

"…we know very little about the risks of continuation therapy (with SSRIs). We have only recently recognised that indefinite maintenance therapy with tricyclics, a group of drugs that we have used for 20 years, increases the risk of sudden death in patients with an arrythmia. When will we know if there are adverse long-term consequences for fluoxetine ?" (Preskorn, 1994)

With benzodiazepines, the most unexpected findings were of lack of evidence of long-term efficacy and the dependence problem. However, the main claim in the UK litigation was that long-term use had brought about cognitive impairment and depersonalisation, in several manifestations. As information coordinator with the Plaintiffs’ legal team, I saw most of this evidence. My impression from memory (1992) is that formal studies provided suggestive rather than conclusive evidence, but there was good deal of clinical suspicion, plus worrying evidence of the very belated recognition of severe behavioural deficits with barbiturates. In law, with a lower burden of proof (51%), it might have gone either way but (for legal reasons and funding problems) the case never came to court.

With antidepressants, the only aspect of behavioural toxicity to have been formally investigated is excessive sedation in the short-term. (Freeman & O’Hanlon, 1992) Apart from unwanted behavioural effects, two possible areas of long-term risk with SSRIs have so far been identified, though their significance is unclear. One concerns often persisting sexual problems, but this has apparently not been investigated. The other concerns the tendency of some SSRIs (mainly paroxetine, fluoxetine and sertraline) to inhibit an important liver enzyme system, increasing the risk of toxicity with many other drugs and of drug interactions. (Nemeroff et al., 1996) Those most at risk are a minority (about 8% in Caucasian populations) whose genetic make-up leads to reduced efficiency in this enzyme system, who are therefore already "poor metabolisers" of the same drugs:

"Poor metabolisers demonstrate longer plasma half-lives and thus, higher steady-state drug concentrations than their ‘fast’ metaboliser counterparts. Accordingly, ‘poor metabolisers’ carry a higher risk for toxicity and/or drug interactions ..." Sindrup and colleagues (1992) reported that "paroxetine reversibly converts normal or extensive metabolisers to the poor metaboliser phenotype. This may be true of several other SSRIs" (Tollefson, 1993)

Uncertainty about long-term risk can be expected to influence some prescribing decisions. This does not appear to have been formally studied, but the dilemmas involved have occasionally been mentioned in published discussions. Some advocate more sparing use, both to avoid over-treatment and because of the possible long-term risks involved:

"My approach is to treat each episode of depression for 6 months and then taper therapy. Some recurrences are as long as 5 years apart. Should patients be committed to indefinite therapy to prevent widely spaced episodes of depression ?" (Preskorn, 1994)

Other experts take the view that intermittent treatment involves a higher risk, notably because of the possibility of strong "rebound" reactions (Roose, 1994) and apparent relapse if treatment is interrupted:

"Some circumstantial evidence suggests that antidepressants are sensitising and increase the risk of recurrence, but without maintenance treatment, patients are going to have a recurring course of illness with devastating consequences. I believe that a decision to start maintenance treatment represents a commitment to long-term therapy, because stopping the medication will lead to recurrence" (Keller, 1994)

Nor have questions about the true nature of relapse, and a possible link with dependence, entirely gone away. In conversation with Healy, one of the pioneers of antidepressant therapy recently mused about this, as follows:

"We are trying to keep people on antidepressants for rather long periods of time and the relapse rate goes up if you stop too soon so you wonder whether... There’s an old article on imipramine in the Canadian Journal of Psychiatry, around the time of the first conference with imipramine in Montreal, saying imipramine is an addictive drug because if you stop it you get depressed again, therefore you are addicted to it. The same model would say that diabetics are addicted to insulin. But there is some truth in it and the question is even more acute with Xanax and panic disorder so I don’t know how it’s going to work out in the long run". (Cole, 1996)

But what does insulin dependence really have to do with the long-term use of antidepressants? The answer in the end comes down to one’s view of the hypothesis that depression is a deficiency disease and that antidepressants work by restoring serotonin to normal levels. Far-fetched as this view of serotonin has to be, (Healy, 1987) it has nevertheless captured professional and public imagination to a remarkable degree. Many doctors and patients need no further persuading and many prospective patients can soon be expected to join in. Early in 1997, the manufacturers of venlafaxine (Effexor/Efexor) began a ‘Direct to Consumer’ advertising campaign in the US (SCRIP, 1997) and that Summer the manufacturers of Prozac followed suit:

"Prozac, the ‘happy pill’, is in the news again, with its makers, Eli Lilly, being criticised for going over doctors’ heads and directly targeting depressed Americans with a big advertising campaign. Two-page colour ads, depicting a dark rain cloud followed by a bright sun, will appear in the US next week in 20 consumer magazines, including Newsweek, Cosmo, Time and Marie Claire, aimed at getting patients to diagnose themselves and then ask their doctor for the drug by name. In Britain, advertising drugs directly to patients is illegal." (Hicks, 1997)

It remains to be seen how long it will take before the public begins to question the fundamental contradiction that arises here: if long-term users of antidepressants are indeed in the position of insulin-dependent diabetics, why have they repeatedly been told that there is no risk of dependence? If the analogy held, antidepressants would have the potential of ‘once on, never off’ type drugs; insulin-dependent diabetics need drugs for life. Related concerns have prompted another pioneer in the field to speculate that serious problems might be looming even now:

"I think the next big issue is going to be the question of long-term treatment of depressive illness. I think what will happen, and it has already begun to happen in the United States, is that patients are going to start suing doctors who haven’t informed them of the course of the illness. There is a general agreement about the course of the illness now - it’s pretty bad - so everyone should be told about it." (Coppen, 1996).

On the other hand, one might question how far the assumption that antidepressants were absolutely not drugs of dependence had coloured understanding of drug action and effectiveness, and the nature and course of depression. This question arises if one rejects the notion that drugs should be regarded almost as nutrients for some frank malnutrition of the mind.

Insulin withdrawal (or shutting down the pancreas) swiftly, dramatically and universally leads to fundamental and quite specific disorders of metabolism. If antidepressants were in some sense drugs of dependence, they would not resemble insulin in this respect. The evidence suggests a much closer link with benzodiazepines:

The thrust of the Defeat Depression Campaign, among many other communications to the general public, has been to say rather the opposite of all this. Perhaps the time has now come to thoroughly investigate what is what and to set the record straight.


Contents page