[Commentary] COMING OFF TRANQUILLIZERS—A SISYPHEAN TOIL

The risks of benzodiazepine use, both short- and long-term, have long been known 1-3. Furthermore, they are undoubtedly drugs associated with problems on attempted reduction in dosage or withdrawal. The discontinuation syndrome can be severe and can prevent the long-term user from ever stopping the medication. Discontinuation is a desirable goal with both anxiolytics and hypnotics 4. It is thus apparent that guidance is needed on efficacious, safe and cost-effective measures to facilitate withdrawal from benzodiazepine use. In this context, the review by Parr et al. is to be welcomed. It is a thorough evaluation of the problem of selecting such regimens. Nevertheless, some caveats can be entered. First, a systematic review, of which this is an excellent example, can be qualitative, describing available trials in terms of their strengths and weaknesses, and their relevance to clinical practice, or quantitative wherein results are combined from various studies 5. However, both may overlook the purpose of an included trial. Not all are primarily efficacy studies but pilot studies or proof-of-concept, and may be underpowered or adopt suboptimal designs. Therefore, some appraisal of the background rationale is helpful. Also, separate appraisal of techniques used to withdraw from anxiolytic as contrasted with hypnotic benzodiazepines is helpful. Adequate follow-up is essential, as relapse is so frequent. As with any therapeutic procedure, the risk/benefit ratio is the crucial variable. Not only is it necessary to establish that a statistically significant difference exists between the trial treatment and the routine treatment or placebo, but the effect size is also crucial. Is the difference of any clinical importance—essentially an informed value judgement? Table 3 is invaluable in setting out both significances and the differences in discontinuation rates. It makes gloomy reading. Simple general practitioner (GP) intervention is not very effective, and more intensive measures help only about half the patients. Of course, these patients are typically those with a poor prognosis, in that previous withdrawal attempts have failed. But what is this comparator ‘routine care’. ‘Routine lack of care’ is sometimes more evident, so that medico-legal considerations are becoming more salient. Many expert GPs are prepared to opine that benzodiazepine treatment extending beyond 2–4 weeks is a breach of duty of care. It may yet be that the law will govern our use of these drugs rather than medical science. The therapeutic landscape has changed in the last few years. The licensing of selective serotonin re-uptake inhibitors (SSRIs) and pregabalin for anxiety disorders and of melatonin agonists for insomnia has opened up the prescribers' therapeutic choices. Accordingly, the use of benzodiazepines can be largely avoided, especially as the newer drugs are generally better tolerated without a risk of dependence or abuse. The authors rightly draw attention to selecting a sequence of treatments and not simply contemporaneous comparisons. A stepped-care approach was suggested by Russell & Lader 6 in which the first two steps involved primary care practitioners—a minimal intervention strategy, then systematic tapering for the failures. Hospital-based schedules then followed but the procedures became more elaborate and consumed more resources. An earlier meta-analysis of available studies suggested that the empirical protocol was supported by the results of controlled clinical trials 7. I would urge the present authors to apply their analytical skills in evaluating the welter of disparate studies to developing this theme of stepped care, so we can give practical evidence-based advice to doctors 8 and patients 9. And not before time! 10. None.