Discontinuation does contribute to relapse
EDITOR-Grof and Schou attempt to explain the discepancies in the evidence about lithium by suggesting that the withdrawal syndrome only occurs in those with more broadly defined bipolar disorder.1 Schou has been an enthusiast for lithium and was involved in an influential double-blind lithium discontinuation study, which had remarkably favourable results for lithium (21 placebo patients relapsed and none on lithium).2 In that study patients were not informed they were in a clinical trial and were deceived into believing they had received tablets from an unreliable batch so that concealed trial medication could be reissued. Although this procedure would now be regarded as unethical because of lack of consent, it does have the advantage that patient expectancies are not created if they remain unaware of the trial. Four placebo patients were excluded from the analysis because for some reason they used some of their old tablets or supplementary lithium from outside sources.
Particular detail was paid to methodological issues because of criticism of the inferences made from previous non-blind trials.3 Patients were matched on total number of previous episodes, and assigned randomly to lithium or placebo, although no details were given of the randomisation procedure. Although it was anticipated that side-effects might compromise the blindness of the trial, and the protocol required that patients with side-effects must be excluded, none were eliminated. Lithium blood results were reported to assessors, but care was reported to have been taken to ensure that fictitious values for the placebo patients approximated to those recorded before the trial and showed variations similar to those found in the patients receiving lithium. No attempt was made to measure the degree of unblinding. Despite the lack of evidence for unblinding or incomplete randomisation it is difficult to believe that bias did not in some way affect the results as they are so strikingly favourable to lithium, particularly in the context of the need to counter criticism at the time and produce an unequivocal answer.
One of the reasons proposed for the striking difference is that all patients had well-diagnosed affective disorders of the endogenous type, the same point about effectiveness as Grof and Schou make about the presence of the withdrawal syndrome. The assumption seems to be that lithium is a specific prophylactic treatment for the biological diseases of manic-depressive and recurrent-depressive disorders. Evidence for lithium withdrawal has not always been found and heightened anxiety, sleep disturbances and irritability have been associated.4 The presence of withdrawal symptoms seems to also be influenced by clinicians and experimenters and not just patients alone. It seems appropriate and timely to propose further research of the implications that drug discontinuation itself contributes to relapse, including the study of nonspecific effects. This is the same conclusion reached by the group, quoted by Grof and Schou, who found that recurrences are greater than before lithium treatment in patients abruptly discontinuing lithium.5
DB Double, Consultant Psychiatrist, Norfolk Mental Health Care NHS Trust, Hellesdon Hospital, Drayton High Road, Norwich NR6 5BE (Duncan_Double@bigfoot.com)
1. Grof P, Schou M. Re: Lithium. http://www.bmj.com/cgi/eletters/316/7141/1330#EL1.
2. Baastrup PC, Poulsen JC, Schou M, Thomsen K, Amdisen A. Prophylactic lithium: Double blind discontinuation in manic-depressive and recurrent-depressive disorders. Lancet 1970;ii:326-330
3. Blackwell B, Shepherd M. Prophylactic lithium: another therapeutic myth? Lancet 1968;i:968-971
4. Balon R, Yeragani VK, Pohl RB, Gershon S. Lithium discontinuation: Withdrawal or relapse? Comp Psychiatry 1988;29:330-334
5. Baldessarini RJ, Tondo L, Viguera AC. Forty years of lithium treatment. Arch Gen Psychiatry 1998;55:93
Competing interests: No competing interests